Primary Storage

Hello,

we got 2 new 5035 FC Storages. We done the initial setup on the first one and add the second to the first. Now we have a two-node cluster.

We create a Pool0 with all SSDs in System1 (Raid6) and a Pool1 with all SSDs in System2.

We download and start the IP Quorum App successful (connect).

Our goal is to create a realtime/live mirrored voulmes.

I create a mirrored Volume. Copy0 to Pool0 and Copy2 to Pool1.

In the details of this volume I see the volume is striped ?!? So if I power off enclosure-1 I got the waring "Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 1 volume."

May be I misunderstand mirrored Volumes ?!?

Now we add the Remote Copy license and modify the system Topology to HyperSwap. Now we got 3 sites. 1=enclosure1, 2=enclosure2, 3=quorum

I create a HyperSwap Volume in Pool0 and Pool1. In the details of the volume, I found again "Virtualization type: Striped".

And same behavior, If I shutdown enclosure1 I got the message "Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 1 volume."

If I shutdown I enclosure2 I got the warning "Powering off the control enclosure will impact system performance." And the volume is online.

But we want the volume to remain accessible on both enclosure.

Please help !

Hello,

we got 2 new 5035 FC Storages. We done the initial setup on the first one and add the second to the first. Now we have a two-node cluster.

We create a Pool0 with all SSDs in System1 (Raid6) and a Pool1 with all SSDs in System2.

We download and start the IP Quorum App successful (connect).

Our goal is to create a realtime/live mirrored voulmes.

I create a mirrored Volume. Copy0 to Pool0 and Copy2 to Pool1.

In the details of this volume I see the volume is striped ?!? So if I power off enclosure-1 I got the waring "Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 1 volume."

May be I misunderstand mirrored Volumes ?!?

Now we add the Remote Copy license and modify the system Topology to HyperSwap. Now we got 3 sites. 1=enclosure1, 2=enclosure2, 3=quorum

I create a HyperSwap Volume in Pool0 and Pool1. In the details of the volume, I found again "Virtualization type: Striped".

And same behavior, If I shutdown enclosure1 I got the message "Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 1 volume."

If I shutdown I enclosure2 I got the warning "Powering off the control enclosure will impact system performance." And the volume is online.

But we want the volume to remain accessible on both enclosure.

Please help !

Hello,

we got 2 new 5035 FC Storages. We done the initial setup on the first one and add the second to the first. Now we have a two-node cluster.

We create a Pool0 with all SSDs in System1 (Raid6) and a Pool1 with all SSDs in System2.

We download and start the IP Quorum App successful (connect).

Our goal is to create a realtime/live mirrored voulmes.

I create a mirrored Volume. Copy0 to Pool0 and Copy2 to Pool1.

In the details of this volume I see the volume is striped ?!? So if I power off enclosure-1 I got the waring "Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 1 volume."

May be I misunderstand mirrored Volumes ?!?

Now we add the Remote Copy license and modify the system Topology to HyperSwap. Now we got 3 sites. 1=enclosure1, 2=enclosure2, 3=quorum

I create a HyperSwap Volume in Pool0 and Pool1. In the details of the volume, I found again "Virtualization type: Striped".

And same behavior, If I shutdown enclosure1 I got the message "Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 1 volume."

If I shutdown I enclosure2 I got the warning "Powering off the control enclosure will impact system performance." And the volume is online.

But we want the volume to remain accessible on both enclosure.

Please help !

Thank you ! Yes you are right a 4-node cluster or 2 enclosure cluster 😊

 
And yes I think the context of a "striped" volume is/was not clear.

For me is a striped volume on 2 or more disks. If I lose one disk the volume is lost:

But now I understand it is a mirrored volume, each "volume copy" will be striped in its respective pool.

By the way what is the difference between Hyperswap and Mirrored Volume ?

.... 

The confusion starts with the GUI message:

After my new knowledge from the red book, I tested the system again and create VVOL hosts. 

While shutdown one enclosure the GUI message is the same:
"Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 2 volumes." 

But now the Host is able to access both volumes. Maybe the GUI Warning is wrong ?

Hello,

we got 2 new 5035 FC Storages. We done the initial setup on the first one and add the second to the first. Now we have a two-node cluster.

We create a Pool0 with all SSDs in System1 (Raid6) and a Pool1 with all SSDs in System2.

We download and start the IP Quorum App successful (connect).

Our goal is to create a realtime/live mirrored voulmes.

I create a mirrored Volume. Copy0 to Pool0 and Copy2 to Pool1.

In the details of this volume I see the volume is striped ?!? So if I power off enclosure-1 I got the waring "Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 1 volume."

May be I misunderstand mirrored Volumes ?!?

Now we add the Remote Copy license and modify the system Topology to HyperSwap. Now we got 3 sites. 1=enclosure1, 2=enclosure2, 3=quorum

I create a HyperSwap Volume in Pool0 and Pool1. In the details of the volume, I found again "Virtualization type: Striped".

And same behavior, If I shutdown enclosure1 I got the message "Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 1 volume."

If I shutdown I enclosure2 I got the warning "Powering off the control enclosure will impact system performance." And the volume is online.

But we want the volume to remain accessible on both enclosure.

Please help !

Thank you ! Yes you are right a 4-node cluster or 2 enclosure cluster 😊

 
And yes I think the context of a "striped" volume is/was not clear.

For me is a striped volume on 2 or more disks. If I lose one disk the volume is lost:

But now I understand it is a mirrored volume, each "volume copy" will be striped in its respective pool.

By the way what is the difference between Hyperswap and Mirrored Volume ?

.... 

The confusion starts with the GUI message:

After my new knowledge from the red book, I tested the system again and create VVOL hosts. 

While shutdown one enclosure the GUI message is the same:
"Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 2 volumes." 

But now the Host is able to access both volumes. Maybe the GUI Warning is wrong ?

Hello,

we got 2 new 5035 FC Storages. We done the initial setup on the first one and add the second to the first. Now we have a two-node cluster.

We create a Pool0 with all SSDs in System1 (Raid6) and a Pool1 with all SSDs in System2.

We download and start the IP Quorum App successful (connect).

Our goal is to create a realtime/live mirrored voulmes.

I create a mirrored Volume. Copy0 to Pool0 and Copy2 to Pool1.

In the details of this volume I see the volume is striped ?!? So if I power off enclosure-1 I got the waring "Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 1 volume."

May be I misunderstand mirrored Volumes ?!?

Now we add the Remote Copy license and modify the system Topology to HyperSwap. Now we got 3 sites. 1=enclosure1, 2=enclosure2, 3=quorum

I create a HyperSwap Volume in Pool0 and Pool1. In the details of the volume, I found again "Virtualization type: Striped".

And same behavior, If I shutdown enclosure1 I got the message "Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 1 volume."

If I shutdown I enclosure2 I got the warning "Powering off the control enclosure will impact system performance." And the volume is online.

But we want the volume to remain accessible on both enclosure.

Please help !

Thank you. We have only 1 enclosure in 2 different rooms. Both connected with 16gbit FC. If I understand you correctly, I didn't need a hyperswap. I will change the system topology. 

Does anybody know why the GUI shows the waring "…the host will not be able to access 2 volumes." 

In my first test, the ESX Host lost the volume. After configuring a VVOL Host, the host can access the volume with ony one enclosure. I think this was a path reconnect issue.

Sorry for my many newbies questions 😊

Thank you ! Yes you are right a 4-node cluster or 2 enclosure cluster 😊

 
And yes I think the context of a "striped" volume is/was not clear.

For me is a striped volume on 2 or more disks. If I lose one disk the volume is lost:

But now I understand it is a mirrored volume, each "volume copy" will be striped in its respective pool.

By the way what is the difference between Hyperswap and Mirrored Volume ?

.... 

The confusion starts with the GUI message:

After my new knowledge from the red book, I tested the system again and create VVOL hosts. 

While shutdown one enclosure the GUI message is the same:
"Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 2 volumes." 

But now the Host is able to access both volumes. Maybe the GUI Warning is wrong ?

Hello,

we got 2 new 5035 FC Storages. We done the initial setup on the first one and add the second to the first. Now we have a two-node cluster.

We create a Pool0 with all SSDs in System1 (Raid6) and a Pool1 with all SSDs in System2.

We download and start the IP Quorum App successful (connect).

Our goal is to create a realtime/live mirrored voulmes.

I create a mirrored Volume. Copy0 to Pool0 and Copy2 to Pool1.

In the details of this volume I see the volume is striped ?!? So if I power off enclosure-1 I got the waring "Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 1 volume."

May be I misunderstand mirrored Volumes ?!?

Now we add the Remote Copy license and modify the system Topology to HyperSwap. Now we got 3 sites. 1=enclosure1, 2=enclosure2, 3=quorum

I create a HyperSwap Volume in Pool0 and Pool1. In the details of the volume, I found again "Virtualization type: Striped".

And same behavior, If I shutdown enclosure1 I got the message "Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 1 volume."

If I shutdown I enclosure2 I got the warning "Powering off the control enclosure will impact system performance." And the volume is online.

But we want the volume to remain accessible on both enclosure.

Please help !

Thank you. We have only 1 enclosure in 2 different rooms. Both connected with 16gbit FC. If I understand you correctly, I didn't need a hyperswap. I will change the system topology. 

Does anybody know why the GUI shows the waring "…the host will not be able to access 2 volumes." 

In my first test, the ESX Host lost the volume. After configuring a VVOL Host, the host can access the volume with ony one enclosure. I think this was a path reconnect issue.

Sorry for my many newbies questions 😊

Thank you ! Yes you are right a 4-node cluster or 2 enclosure cluster 😊

 
And yes I think the context of a "striped" volume is/was not clear.

For me is a striped volume on 2 or more disks. If I lose one disk the volume is lost:

But now I understand it is a mirrored volume, each "volume copy" will be striped in its respective pool.

By the way what is the difference between Hyperswap and Mirrored Volume ?

.... 

The confusion starts with the GUI message:

After my new knowledge from the red book, I tested the system again and create VVOL hosts. 

While shutdown one enclosure the GUI message is the same:
"Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 2 volumes." 

But now the Host is able to access both volumes. Maybe the GUI Warning is wrong ?

Hello,

we got 2 new 5035 FC Storages. We done the initial setup on the first one and add the second to the first. Now we have a two-node cluster.

We create a Pool0 with all SSDs in System1 (Raid6) and a Pool1 with all SSDs in System2.

We download and start the IP Quorum App successful (connect).

Our goal is to create a realtime/live mirrored voulmes.

I create a mirrored Volume. Copy0 to Pool0 and Copy2 to Pool1.

In the details of this volume I see the volume is striped ?!? So if I power off enclosure-1 I got the waring "Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 1 volume."

May be I misunderstand mirrored Volumes ?!?

Now we add the Remote Copy license and modify the system Topology to HyperSwap. Now we got 3 sites. 1=enclosure1, 2=enclosure2, 3=quorum

I create a HyperSwap Volume in Pool0 and Pool1. In the details of the volume, I found again "Virtualization type: Striped".

And same behavior, If I shutdown enclosure1 I got the message "Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 1 volume."

If I shutdown I enclosure2 I got the warning "Powering off the control enclosure will impact system performance." And the volume is online.

But we want the volume to remain accessible on both enclosure.

Please help !

Hello Masteen,

no we have 2 enclosures with each 2 nodes. Site1=enclosure1, Site2=enclosure2, Site3=quorum

The volume should work on both enclosures (4 nodes). So if Site1 is down (enclosure 1) the host(s) should have access to this volume (enclosure 2).

In my screenshot I shutdown enclosure1. The volumes are online in enclosure2, alias site2, alias Bunker:

Thank you. We have only 1 enclosure in 2 different rooms. Both connected with 16gbit FC. If I understand you correctly, I didn't need a hyperswap. I will change the system topology. 

Does anybody know why the GUI shows the waring "…the host will not be able to access 2 volumes." 

In my first test, the ESX Host lost the volume. After configuring a VVOL Host, the host can access the volume with ony one enclosure. I think this was a path reconnect issue.

Sorry for my many newbies questions 😊

Thank you ! Yes you are right a 4-node cluster or 2 enclosure cluster 😊

 
And yes I think the context of a "striped" volume is/was not clear.

For me is a striped volume on 2 or more disks. If I lose one disk the volume is lost:

But now I understand it is a mirrored volume, each "volume copy" will be striped in its respective pool.

By the way what is the difference between Hyperswap and Mirrored Volume ?

.... 

The confusion starts with the GUI message:

After my new knowledge from the red book, I tested the system again and create VVOL hosts. 

While shutdown one enclosure the GUI message is the same:
"Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 2 volumes." 

But now the Host is able to access both volumes. Maybe the GUI Warning is wrong ?

Hello,

we got 2 new 5035 FC Storages. We done the initial setup on the first one and add the second to the first. Now we have a two-node cluster.

We create a Pool0 with all SSDs in System1 (Raid6) and a Pool1 with all SSDs in System2.

We download and start the IP Quorum App successful (connect).

Our goal is to create a realtime/live mirrored voulmes.

I create a mirrored Volume. Copy0 to Pool0 and Copy2 to Pool1.

In the details of this volume I see the volume is striped ?!? So if I power off enclosure-1 I got the waring "Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 1 volume."

May be I misunderstand mirrored Volumes ?!?

Now we add the Remote Copy license and modify the system Topology to HyperSwap. Now we got 3 sites. 1=enclosure1, 2=enclosure2, 3=quorum

I create a HyperSwap Volume in Pool0 and Pool1. In the details of the volume, I found again "Virtualization type: Striped".

And same behavior, If I shutdown enclosure1 I got the message "Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 1 volume."

If I shutdown I enclosure2 I got the warning "Powering off the control enclosure will impact system performance." And the volume is online.

But we want the volume to remain accessible on both enclosure.

Please help !

Hello Masteen,

no we have 2 enclosures with each 2 nodes. Site1=enclosure1, Site2=enclosure2, Site3=quorum

The volume should work on both enclosures (4 nodes). So if Site1 is down (enclosure 1) the host(s) should have access to this volume (enclosure 2).

In my screenshot I shutdown enclosure1. The volumes are online in enclosure2, alias site2, alias Bunker:

Thank you. We have only 1 enclosure in 2 different rooms. Both connected with 16gbit FC. If I understand you correctly, I didn't need a hyperswap. I will change the system topology. 

Does anybody know why the GUI shows the waring "…the host will not be able to access 2 volumes." 

In my first test, the ESX Host lost the volume. After configuring a VVOL Host, the host can access the volume with ony one enclosure. I think this was a path reconnect issue.

Sorry for my many newbies questions 😊

Thank you ! Yes you are right a 4-node cluster or 2 enclosure cluster 😊

 
And yes I think the context of a "striped" volume is/was not clear.

For me is a striped volume on 2 or more disks. If I lose one disk the volume is lost:

But now I understand it is a mirrored volume, each "volume copy" will be striped in its respective pool.

By the way what is the difference between Hyperswap and Mirrored Volume ?

.... 

The confusion starts with the GUI message:

After my new knowledge from the red book, I tested the system again and create VVOL hosts. 

While shutdown one enclosure the GUI message is the same:
"Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 2 volumes." 

But now the Host is able to access both volumes. Maybe the GUI Warning is wrong ?

Hello,

we got 2 new 5035 FC Storages. We done the initial setup on the first one and add the second to the first. Now we have a two-node cluster.

We create a Pool0 with all SSDs in System1 (Raid6) and a Pool1 with all SSDs in System2.

We download and start the IP Quorum App successful (connect).

Our goal is to create a realtime/live mirrored voulmes.

I create a mirrored Volume. Copy0 to Pool0 and Copy2 to Pool1.

In the details of this volume I see the volume is striped ?!? So if I power off enclosure-1 I got the waring "Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 1 volume."

May be I misunderstand mirrored Volumes ?!?

Now we add the Remote Copy license and modify the system Topology to HyperSwap. Now we got 3 sites. 1=enclosure1, 2=enclosure2, 3=quorum

I create a HyperSwap Volume in Pool0 and Pool1. In the details of the volume, I found again "Virtualization type: Striped".

And same behavior, If I shutdown enclosure1 I got the message "Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 1 volume."

If I shutdown I enclosure2 I got the warning "Powering off the control enclosure will impact system performance." And the volume is online.

But we want the volume to remain accessible on both enclosure.

Please help !

Yes I wrote the same:

While shutdown one enclosure the GUI Warning message is:
"Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 2 volumes." 

But the Host is able to access both volumes ?!? Maybe the GUI Warning is wrong ?

And my question is:

Does anybody know why the GUI shows the warning "…the host will not be able to access 2 volumes." 

In my opinion, the warning should not appear. (And with captcha code even more dramatic)

Thank you.

Hello Masteen,

no we have 2 enclosures with each 2 nodes. Site1=enclosure1, Site2=enclosure2, Site3=quorum

The volume should work on both enclosures (4 nodes). So if Site1 is down (enclosure 1) the host(s) should have access to this volume (enclosure 2).

In my screenshot I shutdown enclosure1. The volumes are online in enclosure2, alias site2, alias Bunker:

Thank you. We have only 1 enclosure in 2 different rooms. Both connected with 16gbit FC. If I understand you correctly, I didn't need a hyperswap. I will change the system topology. 

Does anybody know why the GUI shows the waring "…the host will not be able to access 2 volumes." 

In my first test, the ESX Host lost the volume. After configuring a VVOL Host, the host can access the volume with ony one enclosure. I think this was a path reconnect issue.

Sorry for my many newbies questions 😊

Thank you ! Yes you are right a 4-node cluster or 2 enclosure cluster 😊

 
And yes I think the context of a "striped" volume is/was not clear.

For me is a striped volume on 2 or more disks. If I lose one disk the volume is lost:

But now I understand it is a mirrored volume, each "volume copy" will be striped in its respective pool.

By the way what is the difference between Hyperswap and Mirrored Volume ?

.... 

The confusion starts with the GUI message:

After my new knowledge from the red book, I tested the system again and create VVOL hosts. 

While shutdown one enclosure the GUI message is the same:
"Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 2 volumes." 

But now the Host is able to access both volumes. Maybe the GUI Warning is wrong ?

Hello,

we got 2 new 5035 FC Storages. We done the initial setup on the first one and add the second to the first. Now we have a two-node cluster.

We create a Pool0 with all SSDs in System1 (Raid6) and a Pool1 with all SSDs in System2.

We download and start the IP Quorum App successful (connect).

Our goal is to create a realtime/live mirrored voulmes.

I create a mirrored Volume. Copy0 to Pool0 and Copy2 to Pool1.

In the details of this volume I see the volume is striped ?!? So if I power off enclosure-1 I got the waring "Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 1 volume."

May be I misunderstand mirrored Volumes ?!?

Now we add the Remote Copy license and modify the system Topology to HyperSwap. Now we got 3 sites. 1=enclosure1, 2=enclosure2, 3=quorum

I create a HyperSwap Volume in Pool0 and Pool1. In the details of the volume, I found again "Virtualization type: Striped".

And same behavior, If I shutdown enclosure1 I got the message "Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 1 volume."

If I shutdown I enclosure2 I got the warning "Powering off the control enclosure will impact system performance." And the volume is online.

But we want the volume to remain accessible on both enclosure.

Please help !

Yes I wrote the same:

While shutdown one enclosure the GUI Warning message is:
"Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 2 volumes." 

But the Host is able to access both volumes ?!? Maybe the GUI Warning is wrong ?

And my question is:

Does anybody know why the GUI shows the warning "…the host will not be able to access 2 volumes." 

In my opinion, the warning should not appear. (And with captcha code even more dramatic)

Thank you.

Hello Masteen,

no we have 2 enclosures with each 2 nodes. Site1=enclosure1, Site2=enclosure2, Site3=quorum

The volume should work on both enclosures (4 nodes). So if Site1 is down (enclosure 1) the host(s) should have access to this volume (enclosure 2).

In my screenshot I shutdown enclosure1. The volumes are online in enclosure2, alias site2, alias Bunker:

Thank you. We have only 1 enclosure in 2 different rooms. Both connected with 16gbit FC. If I understand you correctly, I didn't need a hyperswap. I will change the system topology. 

Does anybody know why the GUI shows the waring "…the host will not be able to access 2 volumes." 

In my first test, the ESX Host lost the volume. After configuring a VVOL Host, the host can access the volume with ony one enclosure. I think this was a path reconnect issue.

Sorry for my many newbies questions 😊

Thank you ! Yes you are right a 4-node cluster or 2 enclosure cluster 😊

 
And yes I think the context of a "striped" volume is/was not clear.

For me is a striped volume on 2 or more disks. If I lose one disk the volume is lost:

But now I understand it is a mirrored volume, each "volume copy" will be striped in its respective pool.

By the way what is the difference between Hyperswap and Mirrored Volume ?

.... 

The confusion starts with the GUI message:

After my new knowledge from the red book, I tested the system again and create VVOL hosts. 

While shutdown one enclosure the GUI message is the same:
"Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 2 volumes." 

But now the Host is able to access both volumes. Maybe the GUI Warning is wrong ?

Hello,

we got 2 new 5035 FC Storages. We done the initial setup on the first one and add the second to the first. Now we have a two-node cluster.

We create a Pool0 with all SSDs in System1 (Raid6) and a Pool1 with all SSDs in System2.

We download and start the IP Quorum App successful (connect).

Our goal is to create a realtime/live mirrored voulmes.

I create a mirrored Volume. Copy0 to Pool0 and Copy2 to Pool1.

In the details of this volume I see the volume is striped ?!? So if I power off enclosure-1 I got the waring "Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 1 volume."

May be I misunderstand mirrored Volumes ?!?

Now we add the Remote Copy license and modify the system Topology to HyperSwap. Now we got 3 sites. 1=enclosure1, 2=enclosure2, 3=quorum

I create a HyperSwap Volume in Pool0 and Pool1. In the details of the volume, I found again "Virtualization type: Striped".

And same behavior, If I shutdown enclosure1 I got the message "Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 1 volume."

If I shutdown I enclosure2 I got the warning "Powering off the control enclosure will impact system performance." And the volume is online.

But we want the volume to remain accessible on both enclosure.

Please help !

Yes I wrote the same:

While shutdown one enclosure the GUI Warning message is:
"Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 2 volumes." 

But the Host is able to access both volumes ?!? Maybe the GUI Warning is wrong ?

And my question is:

Does anybody know why the GUI shows the warning "…the host will not be able to access 2 volumes." 

In my opinion, the warning should not appear. (And with captcha code even more dramatic)

Thank you.

Hello Masteen,

no we have 2 enclosures with each 2 nodes. Site1=enclosure1, Site2=enclosure2, Site3=quorum

The volume should work on both enclosures (4 nodes). So if Site1 is down (enclosure 1) the host(s) should have access to this volume (enclosure 2).

In my screenshot I shutdown enclosure1. The volumes are online in enclosure2, alias site2, alias Bunker:

Thank you. We have only 1 enclosure in 2 different rooms. Both connected with 16gbit FC. If I understand you correctly, I didn't need a hyperswap. I will change the system topology. 

Does anybody know why the GUI shows the waring "…the host will not be able to access 2 volumes." 

In my first test, the ESX Host lost the volume. After configuring a VVOL Host, the host can access the volume with ony one enclosure. I think this was a path reconnect issue.

Sorry for my many newbies questions 😊

Thank you ! Yes you are right a 4-node cluster or 2 enclosure cluster 😊

 
And yes I think the context of a "striped" volume is/was not clear.

For me is a striped volume on 2 or more disks. If I lose one disk the volume is lost:

But now I understand it is a mirrored volume, each "volume copy" will be striped in its respective pool.

By the way what is the difference between Hyperswap and Mirrored Volume ?

.... 

The confusion starts with the GUI message:

After my new knowledge from the red book, I tested the system again and create VVOL hosts. 

While shutdown one enclosure the GUI message is the same:
"Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 2 volumes." 

But now the Host is able to access both volumes. Maybe the GUI Warning is wrong ?

Hello,

we got 2 new 5035 FC Storages. We done the initial setup on the first one and add the second to the first. Now we have a two-node cluster.

We create a Pool0 with all SSDs in System1 (Raid6) and a Pool1 with all SSDs in System2.

We download and start the IP Quorum App successful (connect).

Our goal is to create a realtime/live mirrored voulmes.

I create a mirrored Volume. Copy0 to Pool0 and Copy2 to Pool1.

In the details of this volume I see the volume is striped ?!? So if I power off enclosure-1 I got the waring "Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 1 volume."

May be I misunderstand mirrored Volumes ?!?

Now we add the Remote Copy license and modify the system Topology to HyperSwap. Now we got 3 sites. 1=enclosure1, 2=enclosure2, 3=quorum

I create a HyperSwap Volume in Pool0 and Pool1. In the details of the volume, I found again "Virtualization type: Striped".

And same behavior, If I shutdown enclosure1 I got the message "Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 1 volume."

If I shutdown I enclosure2 I got the warning "Powering off the control enclosure will impact system performance." And the volume is online.

But we want the volume to remain accessible on both enclosure.

Please help !

"I think that is a generic message as the system is not aware the volumes are mirrored to a pool hosted on a separate system."

It is a pity, Okay thank you!

But tried and tested there is no access to a mirrored volume when the IO Enlosure is shutdown.

So I think we have to use the hyperswap modus to create a high availability mirrored volume.

A voume with automatically change the io group. Is this right ?

Yes I wrote the same:

While shutdown one enclosure the GUI Warning message is:
"Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 2 volumes." 

But the Host is able to access both volumes ?!? Maybe the GUI Warning is wrong ?

And my question is:

Does anybody know why the GUI shows the warning "…the host will not be able to access 2 volumes." 

In my opinion, the warning should not appear. (And with captcha code even more dramatic)

Thank you.

Hello Masteen,

no we have 2 enclosures with each 2 nodes. Site1=enclosure1, Site2=enclosure2, Site3=quorum

The volume should work on both enclosures (4 nodes). So if Site1 is down (enclosure 1) the host(s) should have access to this volume (enclosure 2).

In my screenshot I shutdown enclosure1. The volumes are online in enclosure2, alias site2, alias Bunker:

Thank you. We have only 1 enclosure in 2 different rooms. Both connected with 16gbit FC. If I understand you correctly, I didn't need a hyperswap. I will change the system topology. 

Does anybody know why the GUI shows the waring "…the host will not be able to access 2 volumes." 

In my first test, the ESX Host lost the volume. After configuring a VVOL Host, the host can access the volume with ony one enclosure. I think this was a path reconnect issue.

Sorry for my many newbies questions 😊

Thank you ! Yes you are right a 4-node cluster or 2 enclosure cluster 😊

 
And yes I think the context of a "striped" volume is/was not clear.

For me is a striped volume on 2 or more disks. If I lose one disk the volume is lost:

But now I understand it is a mirrored volume, each "volume copy" will be striped in its respective pool.

By the way what is the difference between Hyperswap and Mirrored Volume ?

.... 

The confusion starts with the GUI message:

After my new knowledge from the red book, I tested the system again and create VVOL hosts. 

While shutdown one enclosure the GUI message is the same:
"Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 2 volumes." 

But now the Host is able to access both volumes. Maybe the GUI Warning is wrong ?

Hello,

we got 2 new 5035 FC Storages. We done the initial setup on the first one and add the second to the first. Now we have a two-node cluster.

We create a Pool0 with all SSDs in System1 (Raid6) and a Pool1 with all SSDs in System2.

We download and start the IP Quorum App successful (connect).

Our goal is to create a realtime/live mirrored voulmes.

I create a mirrored Volume. Copy0 to Pool0 and Copy2 to Pool1.

In the details of this volume I see the volume is striped ?!? So if I power off enclosure-1 I got the waring "Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 1 volume."

May be I misunderstand mirrored Volumes ?!?

Now we add the Remote Copy license and modify the system Topology to HyperSwap. Now we got 3 sites. 1=enclosure1, 2=enclosure2, 3=quorum

I create a HyperSwap Volume in Pool0 and Pool1. In the details of the volume, I found again "Virtualization type: Striped".

And same behavior, If I shutdown enclosure1 I got the message "Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 1 volume."

If I shutdown I enclosure2 I got the warning "Powering off the control enclosure will impact system performance." And the volume is online.

But we want the volume to remain accessible on both enclosure.

Please help !

"I think that is a generic message as the system is not aware the volumes are mirrored to a pool hosted on a separate system."

It is a pity, Okay thank you!

But tried and tested there is no access to a mirrored volume when the IO Enlosure is shutdown.

So I think we have to use the hyperswap modus to create a high availability mirrored volume.

A voume with automatically change the io group. Is this right ?

Yes I wrote the same:

While shutdown one enclosure the GUI Warning message is:
"Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 2 volumes." 

But the Host is able to access both volumes ?!? Maybe the GUI Warning is wrong ?

And my question is:

Does anybody know why the GUI shows the warning "…the host will not be able to access 2 volumes." 

In my opinion, the warning should not appear. (And with captcha code even more dramatic)

Thank you.

Hello Masteen,

no we have 2 enclosures with each 2 nodes. Site1=enclosure1, Site2=enclosure2, Site3=quorum

The volume should work on both enclosures (4 nodes). So if Site1 is down (enclosure 1) the host(s) should have access to this volume (enclosure 2).

In my screenshot I shutdown enclosure1. The volumes are online in enclosure2, alias site2, alias Bunker:

Thank you. We have only 1 enclosure in 2 different rooms. Both connected with 16gbit FC. If I understand you correctly, I didn't need a hyperswap. I will change the system topology. 

Does anybody know why the GUI shows the waring "…the host will not be able to access 2 volumes." 

In my first test, the ESX Host lost the volume. After configuring a VVOL Host, the host can access the volume with ony one enclosure. I think this was a path reconnect issue.

Sorry for my many newbies questions 😊

Thank you ! Yes you are right a 4-node cluster or 2 enclosure cluster 😊

 
And yes I think the context of a "striped" volume is/was not clear.

For me is a striped volume on 2 or more disks. If I lose one disk the volume is lost:

But now I understand it is a mirrored volume, each "volume copy" will be striped in its respective pool.

By the way what is the difference between Hyperswap and Mirrored Volume ?

.... 

The confusion starts with the GUI message:

After my new knowledge from the red book, I tested the system again and create VVOL hosts. 

While shutdown one enclosure the GUI message is the same:
"Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 2 volumes." 

But now the Host is able to access both volumes. Maybe the GUI Warning is wrong ?

Hello,

we got 2 new 5035 FC Storages. We done the initial setup on the first one and add the second to the first. Now we have a two-node cluster.

We create a Pool0 with all SSDs in System1 (Raid6) and a Pool1 with all SSDs in System2.

We download and start the IP Quorum App successful (connect).

Our goal is to create a realtime/live mirrored voulmes.

I create a mirrored Volume. Copy0 to Pool0 and Copy2 to Pool1.

In the details of this volume I see the volume is striped ?!? So if I power off enclosure-1 I got the waring "Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 1 volume."

May be I misunderstand mirrored Volumes ?!?

Now we add the Remote Copy license and modify the system Topology to HyperSwap. Now we got 3 sites. 1=enclosure1, 2=enclosure2, 3=quorum

I create a HyperSwap Volume in Pool0 and Pool1. In the details of the volume, I found again "Virtualization type: Striped".

And same behavior, If I shutdown enclosure1 I got the message "Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 1 volume."

If I shutdown I enclosure2 I got the warning "Powering off the control enclosure will impact system performance." And the volume is online.

But we want the volume to remain accessible on both enclosure.

Please help !

"I think that is a generic message as the system is not aware the volumes are mirrored to a pool hosted on a separate system."

It is a pity, Okay thank you!

But tried and tested there is no access to a mirrored volume when the IO Enlosure is shutdown.

So I think we have to use the hyperswap modus to create a high availability mirrored volume.

A voume with automatically change the io group. Is this right ?

Yes I wrote the same:

While shutdown one enclosure the GUI Warning message is:
"Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 2 volumes." 

But the Host is able to access both volumes ?!? Maybe the GUI Warning is wrong ?

And my question is:

Does anybody know why the GUI shows the warning "…the host will not be able to access 2 volumes." 

In my opinion, the warning should not appear. (And with captcha code even more dramatic)

Thank you.

Hello Masteen,

no we have 2 enclosures with each 2 nodes. Site1=enclosure1, Site2=enclosure2, Site3=quorum

The volume should work on both enclosures (4 nodes). So if Site1 is down (enclosure 1) the host(s) should have access to this volume (enclosure 2).

In my screenshot I shutdown enclosure1. The volumes are online in enclosure2, alias site2, alias Bunker:

Thank you. We have only 1 enclosure in 2 different rooms. Both connected with 16gbit FC. If I understand you correctly, I didn't need a hyperswap. I will change the system topology. 

Does anybody know why the GUI shows the waring "…the host will not be able to access 2 volumes." 

In my first test, the ESX Host lost the volume. After configuring a VVOL Host, the host can access the volume with ony one enclosure. I think this was a path reconnect issue.

Sorry for my many newbies questions 😊

Thank you ! Yes you are right a 4-node cluster or 2 enclosure cluster 😊

 
And yes I think the context of a "striped" volume is/was not clear.

For me is a striped volume on 2 or more disks. If I lose one disk the volume is lost:

But now I understand it is a mirrored volume, each "volume copy" will be striped in its respective pool.

By the way what is the difference between Hyperswap and Mirrored Volume ?

.... 

The confusion starts with the GUI message:

After my new knowledge from the red book, I tested the system again and create VVOL hosts. 

While shutdown one enclosure the GUI message is the same:
"Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 2 volumes." 

But now the Host is able to access both volumes. Maybe the GUI Warning is wrong ?

Hello,

we got 2 new 5035 FC Storages. We done the initial setup on the first one and add the second to the first. Now we have a two-node cluster.

We create a Pool0 with all SSDs in System1 (Raid6) and a Pool1 with all SSDs in System2.

We download and start the IP Quorum App successful (connect).

Our goal is to create a realtime/live mirrored voulmes.

I create a mirrored Volume. Copy0 to Pool0 and Copy2 to Pool1.

In the details of this volume I see the volume is striped ?!? So if I power off enclosure-1 I got the waring "Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 1 volume."

May be I misunderstand mirrored Volumes ?!?

Now we add the Remote Copy license and modify the system Topology to HyperSwap. Now we got 3 sites. 1=enclosure1, 2=enclosure2, 3=quorum

I create a HyperSwap Volume in Pool0 and Pool1. In the details of the volume, I found again "Virtualization type: Striped".

And same behavior, If I shutdown enclosure1 I got the message "Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 1 volume."

If I shutdown I enclosure2 I got the warning "Powering off the control enclosure will impact system performance." And the volume is online.

But we want the volume to remain accessible on both enclosure.

Please help !

Hallo Patrik,

vielen vielen Dank für die tolle Erklärung !!! 😊

Dann ist es klar, HyperSwap muss für unseren HA-Betrieb genutzt werden. Den hier soll niemand manuell eingreifen.

Ist der Performance Unterschied zwischen CDM 8.5.1.x signifikant zum LTS 8.5.0.5? Da so ein SAN-System in der Regel nicht so häufig gewartet wird (FW-Updates), denke ich LTS könnte das bessere für uns sein.

Ich melde mich hier mit den Ergebnissen nach FW-Update und der HyperSwap Config.

PS: Woher war meine Muttersprache ersichtlich? 😊

"I think that is a generic message as the system is not aware the volumes are mirrored to a pool hosted on a separate system."

It is a pity, Okay thank you!

But tried and tested there is no access to a mirrored volume when the IO Enlosure is shutdown.

So I think we have to use the hyperswap modus to create a high availability mirrored volume.

A voume with automatically change the io group. Is this right ?

Yes I wrote the same:

While shutdown one enclosure the GUI Warning message is:
"Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 2 volumes." 

But the Host is able to access both volumes ?!? Maybe the GUI Warning is wrong ?

And my question is:

Does anybody know why the GUI shows the warning "…the host will not be able to access 2 volumes." 

In my opinion, the warning should not appear. (And with captcha code even more dramatic)

Thank you.

Hello Masteen,

no we have 2 enclosures with each 2 nodes. Site1=enclosure1, Site2=enclosure2, Site3=quorum

The volume should work on both enclosures (4 nodes). So if Site1 is down (enclosure 1) the host(s) should have access to this volume (enclosure 2).

In my screenshot I shutdown enclosure1. The volumes are online in enclosure2, alias site2, alias Bunker:

Thank you. We have only 1 enclosure in 2 different rooms. Both connected with 16gbit FC. If I understand you correctly, I didn't need a hyperswap. I will change the system topology. 

Does anybody know why the GUI shows the waring "…the host will not be able to access 2 volumes." 

In my first test, the ESX Host lost the volume. After configuring a VVOL Host, the host can access the volume with ony one enclosure. I think this was a path reconnect issue.

Sorry for my many newbies questions 😊

Thank you ! Yes you are right a 4-node cluster or 2 enclosure cluster 😊

 
And yes I think the context of a "striped" volume is/was not clear.

For me is a striped volume on 2 or more disks. If I lose one disk the volume is lost:

But now I understand it is a mirrored volume, each "volume copy" will be striped in its respective pool.

By the way what is the difference between Hyperswap and Mirrored Volume ?

.... 

The confusion starts with the GUI message:

After my new knowledge from the red book, I tested the system again and create VVOL hosts. 

While shutdown one enclosure the GUI message is the same:
"Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 2 volumes." 

But now the Host is able to access both volumes. Maybe the GUI Warning is wrong ?

Hello,

we got 2 new 5035 FC Storages. We done the initial setup on the first one and add the second to the first. Now we have a two-node cluster.

We create a Pool0 with all SSDs in System1 (Raid6) and a Pool1 with all SSDs in System2.

We download and start the IP Quorum App successful (connect).

Our goal is to create a realtime/live mirrored voulmes.

I create a mirrored Volume. Copy0 to Pool0 and Copy2 to Pool1.

In the details of this volume I see the volume is striped ?!? So if I power off enclosure-1 I got the waring "Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 1 volume."

May be I misunderstand mirrored Volumes ?!?

Now we add the Remote Copy license and modify the system Topology to HyperSwap. Now we got 3 sites. 1=enclosure1, 2=enclosure2, 3=quorum

I create a HyperSwap Volume in Pool0 and Pool1. In the details of the volume, I found again "Virtualization type: Striped".

And same behavior, If I shutdown enclosure1 I got the message "Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 1 volume."

If I shutdown I enclosure2 I got the warning "Powering off the control enclosure will impact system performance." And the volume is online.

But we want the volume to remain accessible on both enclosure.

Please help !

Finale Rückmeldung:

Es funktioniert mit Hyperswap 😊

Das Volume ist nun auf beiden IO Gruppen zugänglich:

Das Enclosure1 (mit Node1+2) als auch das Enclosure2 (mit Node 3+4) kann problemlos heruntergefahren oder im laufenden Betrieb getrennt werden.

Die aktiv arbeitende VM auf dem Volume bekommt davon nichts mit. Auch die Reconnects nach dem booten der Nodes funktioniert schnell und zuverlässig (Round Robin).

Leider gibt es immer noch die blöde Meldung in der IBM GUI, dass das Volume ohne Enclosure-2 nicht mehr verfügbar sein soll. Das ist aber nachweislich falsch:

Ich hoffe das hier IBM später man nachbessert.

Mit je 6x 1,92TB SAS SSD (Raid6) erreichen die beiden 5035 übrigens über 60.000 IOPS:

32 Outstanding I/Os; Default Access (100% Random, 67% Read, 2K), IBM LTS 8.5.0.5

------------------------------------------------------------------------------------------------------------------

Final feedback:

It works with Hyperswap 😊

The volume is now accessible on both IO groups:

Enclosure1 (with Node1+2) as well as Enclosure2 (with Node 3+4) can be powered down or hot disconnected. A working VM on the volume does not stop.

Unfortunately there is still the stupid message in the IBM GUI:
But this is wrong.


I hope that IBM will correct this later.

The system with 6x 1.92TB SAS SSD (Raid6) in each enclousre, reached over 60,000 IOPS:

32 outstanding I/Os; Default Access (100% Random, 67% Read, 2K), IBM LTS 8.5.0.5

Hallo Patrik,

vielen vielen Dank für die tolle Erklärung !!! 😊

Dann ist es klar, HyperSwap muss für unseren HA-Betrieb genutzt werden. Den hier soll niemand manuell eingreifen.

Ist der Performance Unterschied zwischen CDM 8.5.1.x signifikant zum LTS 8.5.0.5? Da so ein SAN-System in der Regel nicht so häufig gewartet wird (FW-Updates), denke ich LTS könnte das bessere für uns sein.

Ich melde mich hier mit den Ergebnissen nach FW-Update und der HyperSwap Config.

PS: Woher war meine Muttersprache ersichtlich? 😊

"I think that is a generic message as the system is not aware the volumes are mirrored to a pool hosted on a separate system."

It is a pity, Okay thank you!

But tried and tested there is no access to a mirrored volume when the IO Enlosure is shutdown.

So I think we have to use the hyperswap modus to create a high availability mirrored volume.

A voume with automatically change the io group. Is this right ?

Yes I wrote the same:

While shutdown one enclosure the GUI Warning message is:
"Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 2 volumes." 

But the Host is able to access both volumes ?!? Maybe the GUI Warning is wrong ?

And my question is:

Does anybody know why the GUI shows the warning "…the host will not be able to access 2 volumes." 

In my opinion, the warning should not appear. (And with captcha code even more dramatic)

Thank you.

Hello Masteen,

no we have 2 enclosures with each 2 nodes. Site1=enclosure1, Site2=enclosure2, Site3=quorum

The volume should work on both enclosures (4 nodes). So if Site1 is down (enclosure 1) the host(s) should have access to this volume (enclosure 2).

In my screenshot I shutdown enclosure1. The volumes are online in enclosure2, alias site2, alias Bunker:

Thank you. We have only 1 enclosure in 2 different rooms. Both connected with 16gbit FC. If I understand you correctly, I didn't need a hyperswap. I will change the system topology. 

Does anybody know why the GUI shows the waring "…the host will not be able to access 2 volumes." 

In my first test, the ESX Host lost the volume. After configuring a VVOL Host, the host can access the volume with ony one enclosure. I think this was a path reconnect issue.

Sorry for my many newbies questions 😊

Thank you ! Yes you are right a 4-node cluster or 2 enclosure cluster 😊

 
And yes I think the context of a "striped" volume is/was not clear.

For me is a striped volume on 2 or more disks. If I lose one disk the volume is lost:

But now I understand it is a mirrored volume, each "volume copy" will be striped in its respective pool.

By the way what is the difference between Hyperswap and Mirrored Volume ?

.... 

The confusion starts with the GUI message:

After my new knowledge from the red book, I tested the system again and create VVOL hosts. 

While shutdown one enclosure the GUI message is the same:
"Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 2 volumes." 

But now the Host is able to access both volumes. Maybe the GUI Warning is wrong ?

Hello,

we got 2 new 5035 FC Storages. We done the initial setup on the first one and add the second to the first. Now we have a two-node cluster.

We create a Pool0 with all SSDs in System1 (Raid6) and a Pool1 with all SSDs in System2.

We download and start the IP Quorum App successful (connect).

Our goal is to create a realtime/live mirrored voulmes.

I create a mirrored Volume. Copy0 to Pool0 and Copy2 to Pool1.

In the details of this volume I see the volume is striped ?!? So if I power off enclosure-1 I got the waring "Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 1 volume."

May be I misunderstand mirrored Volumes ?!?

Now we add the Remote Copy license and modify the system Topology to HyperSwap. Now we got 3 sites. 1=enclosure1, 2=enclosure2, 3=quorum

I create a HyperSwap Volume in Pool0 and Pool1. In the details of the volume, I found again "Virtualization type: Striped".

And same behavior, If I shutdown enclosure1 I got the message "Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 1 volume."

If I shutdown I enclosure2 I got the warning "Powering off the control enclosure will impact system performance." And the volume is online.

But we want the volume to remain accessible on both enclosure.

Please help !

Finale Rückmeldung:

Es funktioniert mit Hyperswap 😊

Das Volume ist nun auf beiden IO Gruppen zugänglich:

Das Enclosure1 (mit Node1+2) als auch das Enclosure2 (mit Node 3+4) kann problemlos heruntergefahren oder im laufenden Betrieb getrennt werden.

Die aktiv arbeitende VM auf dem Volume bekommt davon nichts mit. Auch die Reconnects nach dem booten der Nodes funktioniert schnell und zuverlässig (Round Robin).

Leider gibt es immer noch die blöde Meldung in der IBM GUI, dass das Volume ohne Enclosure-2 nicht mehr verfügbar sein soll. Das ist aber nachweislich falsch:

Ich hoffe das hier IBM später man nachbessert.

Mit je 6x 1,92TB SAS SSD (Raid6) erreichen die beiden 5035 übrigens über 60.000 IOPS:

32 Outstanding I/Os; Default Access (100% Random, 67% Read, 2K), IBM LTS 8.5.0.5

------------------------------------------------------------------------------------------------------------------

Final feedback:

It works with Hyperswap 😊

The volume is now accessible on both IO groups:

Enclosure1 (with Node1+2) as well as Enclosure2 (with Node 3+4) can be powered down or hot disconnected. A working VM on the volume does not stop.

Unfortunately there is still the stupid message in the IBM GUI:
But this is wrong.


I hope that IBM will correct this later.

The system with 6x 1.92TB SAS SSD (Raid6) in each enclousre, reached over 60,000 IOPS:

32 outstanding I/Os; Default Access (100% Random, 67% Read, 2K), IBM LTS 8.5.0.5

Hallo Patrik,

vielen vielen Dank für die tolle Erklärung !!! 😊

Dann ist es klar, HyperSwap muss für unseren HA-Betrieb genutzt werden. Den hier soll niemand manuell eingreifen.

Ist der Performance Unterschied zwischen CDM 8.5.1.x signifikant zum LTS 8.5.0.5? Da so ein SAN-System in der Regel nicht so häufig gewartet wird (FW-Updates), denke ich LTS könnte das bessere für uns sein.

Ich melde mich hier mit den Ergebnissen nach FW-Update und der HyperSwap Config.

PS: Woher war meine Muttersprache ersichtlich? 😊

"I think that is a generic message as the system is not aware the volumes are mirrored to a pool hosted on a separate system."

It is a pity, Okay thank you!

But tried and tested there is no access to a mirrored volume when the IO Enlosure is shutdown.

So I think we have to use the hyperswap modus to create a high availability mirrored volume.

A voume with automatically change the io group. Is this right ?

Yes I wrote the same:

While shutdown one enclosure the GUI Warning message is:
"Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 2 volumes." 

But the Host is able to access both volumes ?!? Maybe the GUI Warning is wrong ?

And my question is:

Does anybody know why the GUI shows the warning "…the host will not be able to access 2 volumes." 

In my opinion, the warning should not appear. (And with captcha code even more dramatic)

Thank you.

Hello Masteen,

no we have 2 enclosures with each 2 nodes. Site1=enclosure1, Site2=enclosure2, Site3=quorum

The volume should work on both enclosures (4 nodes). So if Site1 is down (enclosure 1) the host(s) should have access to this volume (enclosure 2).

In my screenshot I shutdown enclosure1. The volumes are online in enclosure2, alias site2, alias Bunker:

Thank you. We have only 1 enclosure in 2 different rooms. Both connected with 16gbit FC. If I understand you correctly, I didn't need a hyperswap. I will change the system topology. 

Does anybody know why the GUI shows the waring "…the host will not be able to access 2 volumes." 

In my first test, the ESX Host lost the volume. After configuring a VVOL Host, the host can access the volume with ony one enclosure. I think this was a path reconnect issue.

Sorry for my many newbies questions 😊

Thank you ! Yes you are right a 4-node cluster or 2 enclosure cluster 😊

 
And yes I think the context of a "striped" volume is/was not clear.

For me is a striped volume on 2 or more disks. If I lose one disk the volume is lost:

But now I understand it is a mirrored volume, each "volume copy" will be striped in its respective pool.

By the way what is the difference between Hyperswap and Mirrored Volume ?

.... 

The confusion starts with the GUI message:

After my new knowledge from the red book, I tested the system again and create VVOL hosts. 

While shutdown one enclosure the GUI message is the same:
"Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 2 volumes." 

But now the Host is able to access both volumes. Maybe the GUI Warning is wrong ?

Hello,

we got 2 new 5035 FC Storages. We done the initial setup on the first one and add the second to the first. Now we have a two-node cluster.

We create a Pool0 with all SSDs in System1 (Raid6) and a Pool1 with all SSDs in System2.

We download and start the IP Quorum App successful (connect).

Our goal is to create a realtime/live mirrored voulmes.

I create a mirrored Volume. Copy0 to Pool0 and Copy2 to Pool1.

In the details of this volume I see the volume is striped ?!? So if I power off enclosure-1 I got the waring "Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 1 volume."

May be I misunderstand mirrored Volumes ?!?

Now we add the Remote Copy license and modify the system Topology to HyperSwap. Now we got 3 sites. 1=enclosure1, 2=enclosure2, 3=quorum

I create a HyperSwap Volume in Pool0 and Pool1. In the details of the volume, I found again "Virtualization type: Striped".

And same behavior, If I shutdown enclosure1 I got the message "Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 1 volume."

If I shutdown I enclosure2 I got the warning "Powering off the control enclosure will impact system performance." And the volume is online.

But we want the volume to remain accessible on both enclosure.

Please help !

In my test environment unfortunately no. I must set the copy-volume manual online.

Of course, you can script it. But now VMWare show a new Datastore with a new ID.

This datastore (volume) have the same content, but it is a new volume.

Therefore, have to use hyperswap.

Finale Rückmeldung:

Es funktioniert mit Hyperswap 😊

Das Volume ist nun auf beiden IO Gruppen zugänglich:

Das Enclosure1 (mit Node1+2) als auch das Enclosure2 (mit Node 3+4) kann problemlos heruntergefahren oder im laufenden Betrieb getrennt werden.

Die aktiv arbeitende VM auf dem Volume bekommt davon nichts mit. Auch die Reconnects nach dem booten der Nodes funktioniert schnell und zuverlässig (Round Robin).

Leider gibt es immer noch die blöde Meldung in der IBM GUI, dass das Volume ohne Enclosure-2 nicht mehr verfügbar sein soll. Das ist aber nachweislich falsch:

Ich hoffe das hier IBM später man nachbessert.

Mit je 6x 1,92TB SAS SSD (Raid6) erreichen die beiden 5035 übrigens über 60.000 IOPS:

32 Outstanding I/Os; Default Access (100% Random, 67% Read, 2K), IBM LTS 8.5.0.5

------------------------------------------------------------------------------------------------------------------

Final feedback:

It works with Hyperswap 😊

The volume is now accessible on both IO groups:

Enclosure1 (with Node1+2) as well as Enclosure2 (with Node 3+4) can be powered down or hot disconnected. A working VM on the volume does not stop.

Unfortunately there is still the stupid message in the IBM GUI:
But this is wrong.


I hope that IBM will correct this later.

The system with 6x 1.92TB SAS SSD (Raid6) in each enclousre, reached over 60,000 IOPS:

32 outstanding I/Os; Default Access (100% Random, 67% Read, 2K), IBM LTS 8.5.0.5

Hallo Patrik,

vielen vielen Dank für die tolle Erklärung !!! 😊

Dann ist es klar, HyperSwap muss für unseren HA-Betrieb genutzt werden. Den hier soll niemand manuell eingreifen.

Ist der Performance Unterschied zwischen CDM 8.5.1.x signifikant zum LTS 8.5.0.5? Da so ein SAN-System in der Regel nicht so häufig gewartet wird (FW-Updates), denke ich LTS könnte das bessere für uns sein.

Ich melde mich hier mit den Ergebnissen nach FW-Update und der HyperSwap Config.

PS: Woher war meine Muttersprache ersichtlich? 😊

"I think that is a generic message as the system is not aware the volumes are mirrored to a pool hosted on a separate system."

It is a pity, Okay thank you!

But tried and tested there is no access to a mirrored volume when the IO Enlosure is shutdown.

So I think we have to use the hyperswap modus to create a high availability mirrored volume.

A voume with automatically change the io group. Is this right ?

Yes I wrote the same:

While shutdown one enclosure the GUI Warning message is:
"Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 2 volumes." 

But the Host is able to access both volumes ?!? Maybe the GUI Warning is wrong ?

And my question is:

Does anybody know why the GUI shows the warning "…the host will not be able to access 2 volumes." 

In my opinion, the warning should not appear. (And with captcha code even more dramatic)

Thank you.

Hello Masteen,

no we have 2 enclosures with each 2 nodes. Site1=enclosure1, Site2=enclosure2, Site3=quorum

The volume should work on both enclosures (4 nodes). So if Site1 is down (enclosure 1) the host(s) should have access to this volume (enclosure 2).

In my screenshot I shutdown enclosure1. The volumes are online in enclosure2, alias site2, alias Bunker:

Thank you. We have only 1 enclosure in 2 different rooms. Both connected with 16gbit FC. If I understand you correctly, I didn't need a hyperswap. I will change the system topology. 

Does anybody know why the GUI shows the waring "…the host will not be able to access 2 volumes." 

In my first test, the ESX Host lost the volume. After configuring a VVOL Host, the host can access the volume with ony one enclosure. I think this was a path reconnect issue.

Sorry for my many newbies questions 😊

Thank you ! Yes you are right a 4-node cluster or 2 enclosure cluster 😊

 
And yes I think the context of a "striped" volume is/was not clear.

For me is a striped volume on 2 or more disks. If I lose one disk the volume is lost:

But now I understand it is a mirrored volume, each "volume copy" will be striped in its respective pool.

By the way what is the difference between Hyperswap and Mirrored Volume ?

.... 

The confusion starts with the GUI message:

After my new knowledge from the red book, I tested the system again and create VVOL hosts. 

While shutdown one enclosure the GUI message is the same:
"Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 2 volumes." 

But now the Host is able to access both volumes. Maybe the GUI Warning is wrong ?

Hello,

we got 2 new 5035 FC Storages. We done the initial setup on the first one and add the second to the first. Now we have a two-node cluster.

We create a Pool0 with all SSDs in System1 (Raid6) and a Pool1 with all SSDs in System2.

We download and start the IP Quorum App successful (connect).

Our goal is to create a realtime/live mirrored voulmes.

I create a mirrored Volume. Copy0 to Pool0 and Copy2 to Pool1.

In the details of this volume I see the volume is striped ?!? So if I power off enclosure-1 I got the waring "Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 1 volume."

May be I misunderstand mirrored Volumes ?!?

Now we add the Remote Copy license and modify the system Topology to HyperSwap. Now we got 3 sites. 1=enclosure1, 2=enclosure2, 3=quorum

I create a HyperSwap Volume in Pool0 and Pool1. In the details of the volume, I found again "Virtualization type: Striped".

And same behavior, If I shutdown enclosure1 I got the message "Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 1 volume."

If I shutdown I enclosure2 I got the warning "Powering off the control enclosure will impact system performance." And the volume is online.

But we want the volume to remain accessible on both enclosure.

Please help !

In my test environment unfortunately no. I must set the copy-volume manual online.

Of course, you can script it. But now VMWare show a new Datastore with a new ID.

This datastore (volume) have the same content, but it is a new volume.

Therefore, have to use hyperswap.

Finale Rückmeldung:

Es funktioniert mit Hyperswap 😊

Das Volume ist nun auf beiden IO Gruppen zugänglich:

Das Enclosure1 (mit Node1+2) als auch das Enclosure2 (mit Node 3+4) kann problemlos heruntergefahren oder im laufenden Betrieb getrennt werden.

Die aktiv arbeitende VM auf dem Volume bekommt davon nichts mit. Auch die Reconnects nach dem booten der Nodes funktioniert schnell und zuverlässig (Round Robin).

Leider gibt es immer noch die blöde Meldung in der IBM GUI, dass das Volume ohne Enclosure-2 nicht mehr verfügbar sein soll. Das ist aber nachweislich falsch:

Ich hoffe das hier IBM später man nachbessert.

Mit je 6x 1,92TB SAS SSD (Raid6) erreichen die beiden 5035 übrigens über 60.000 IOPS:

32 Outstanding I/Os; Default Access (100% Random, 67% Read, 2K), IBM LTS 8.5.0.5

------------------------------------------------------------------------------------------------------------------

Final feedback:

It works with Hyperswap 😊

The volume is now accessible on both IO groups:

Enclosure1 (with Node1+2) as well as Enclosure2 (with Node 3+4) can be powered down or hot disconnected. A working VM on the volume does not stop.

Unfortunately there is still the stupid message in the IBM GUI:
But this is wrong.


I hope that IBM will correct this later.

The system with 6x 1.92TB SAS SSD (Raid6) in each enclousre, reached over 60,000 IOPS:

32 outstanding I/Os; Default Access (100% Random, 67% Read, 2K), IBM LTS 8.5.0.5

Hallo Patrik,

vielen vielen Dank für die tolle Erklärung !!! 😊

Dann ist es klar, HyperSwap muss für unseren HA-Betrieb genutzt werden. Den hier soll niemand manuell eingreifen.

Ist der Performance Unterschied zwischen CDM 8.5.1.x signifikant zum LTS 8.5.0.5? Da so ein SAN-System in der Regel nicht so häufig gewartet wird (FW-Updates), denke ich LTS könnte das bessere für uns sein.

Ich melde mich hier mit den Ergebnissen nach FW-Update und der HyperSwap Config.

PS: Woher war meine Muttersprache ersichtlich? 😊

"I think that is a generic message as the system is not aware the volumes are mirrored to a pool hosted on a separate system."

It is a pity, Okay thank you!

But tried and tested there is no access to a mirrored volume when the IO Enlosure is shutdown.

So I think we have to use the hyperswap modus to create a high availability mirrored volume.

A voume with automatically change the io group. Is this right ?

Yes I wrote the same:

While shutdown one enclosure the GUI Warning message is:
"Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 2 volumes." 

But the Host is able to access both volumes ?!? Maybe the GUI Warning is wrong ?

And my question is:

Does anybody know why the GUI shows the warning "…the host will not be able to access 2 volumes." 

In my opinion, the warning should not appear. (And with captcha code even more dramatic)

Thank you.

Hello Masteen,

no we have 2 enclosures with each 2 nodes. Site1=enclosure1, Site2=enclosure2, Site3=quorum

The volume should work on both enclosures (4 nodes). So if Site1 is down (enclosure 1) the host(s) should have access to this volume (enclosure 2).

In my screenshot I shutdown enclosure1. The volumes are online in enclosure2, alias site2, alias Bunker:

Thank you. We have only 1 enclosure in 2 different rooms. Both connected with 16gbit FC. If I understand you correctly, I didn't need a hyperswap. I will change the system topology. 

Does anybody know why the GUI shows the waring "…the host will not be able to access 2 volumes." 

In my first test, the ESX Host lost the volume. After configuring a VVOL Host, the host can access the volume with ony one enclosure. I think this was a path reconnect issue.

Sorry for my many newbies questions 😊

Thank you ! Yes you are right a 4-node cluster or 2 enclosure cluster 😊

 
And yes I think the context of a "striped" volume is/was not clear.

For me is a striped volume on 2 or more disks. If I lose one disk the volume is lost:

But now I understand it is a mirrored volume, each "volume copy" will be striped in its respective pool.

By the way what is the difference between Hyperswap and Mirrored Volume ?

.... 

The confusion starts with the GUI message:

After my new knowledge from the red book, I tested the system again and create VVOL hosts. 

While shutdown one enclosure the GUI message is the same:
"Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 2 volumes." 

But now the Host is able to access both volumes. Maybe the GUI Warning is wrong ?

Hello,

we got 2 new 5035 FC Storages. We done the initial setup on the first one and add the second to the first. Now we have a two-node cluster.

We create a Pool0 with all SSDs in System1 (Raid6) and a Pool1 with all SSDs in System2.

We download and start the IP Quorum App successful (connect).

Our goal is to create a realtime/live mirrored voulmes.

I create a mirrored Volume. Copy0 to Pool0 and Copy2 to Pool1.

In the details of this volume I see the volume is striped ?!? So if I power off enclosure-1 I got the waring "Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 1 volume."

May be I misunderstand mirrored Volumes ?!?

Now we add the Remote Copy license and modify the system Topology to HyperSwap. Now we got 3 sites. 1=enclosure1, 2=enclosure2, 3=quorum

I create a HyperSwap Volume in Pool0 and Pool1. In the details of the volume, I found again "Virtualization type: Striped".

And same behavior, If I shutdown enclosure1 I got the message "Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 1 volume."

If I shutdown I enclosure2 I got the warning "Powering off the control enclosure will impact system performance." And the volume is online.

But we want the volume to remain accessible on both enclosure.

Please help !

In my test environment unfortunately no. I must set the copy-volume manual online.

Of course, you can script it. But now VMWare show a new Datastore with a new ID.

This datastore (volume) have the same content, but it is a new volume.

Therefore, have to use hyperswap.

Finale Rückmeldung:

Es funktioniert mit Hyperswap 😊

Das Volume ist nun auf beiden IO Gruppen zugänglich:

Das Enclosure1 (mit Node1+2) als auch das Enclosure2 (mit Node 3+4) kann problemlos heruntergefahren oder im laufenden Betrieb getrennt werden.

Die aktiv arbeitende VM auf dem Volume bekommt davon nichts mit. Auch die Reconnects nach dem booten der Nodes funktioniert schnell und zuverlässig (Round Robin).

Leider gibt es immer noch die blöde Meldung in der IBM GUI, dass das Volume ohne Enclosure-2 nicht mehr verfügbar sein soll. Das ist aber nachweislich falsch:

Ich hoffe das hier IBM später man nachbessert.

Mit je 6x 1,92TB SAS SSD (Raid6) erreichen die beiden 5035 übrigens über 60.000 IOPS:

32 Outstanding I/Os; Default Access (100% Random, 67% Read, 2K), IBM LTS 8.5.0.5

------------------------------------------------------------------------------------------------------------------

Final feedback:

It works with Hyperswap 😊

The volume is now accessible on both IO groups:

Enclosure1 (with Node1+2) as well as Enclosure2 (with Node 3+4) can be powered down or hot disconnected. A working VM on the volume does not stop.

Unfortunately there is still the stupid message in the IBM GUI:
But this is wrong.


I hope that IBM will correct this later.

The system with 6x 1.92TB SAS SSD (Raid6) in each enclousre, reached over 60,000 IOPS:

32 outstanding I/Os; Default Access (100% Random, 67% Read, 2K), IBM LTS 8.5.0.5

Hallo Patrik,

vielen vielen Dank für die tolle Erklärung !!! 😊

Dann ist es klar, HyperSwap muss für unseren HA-Betrieb genutzt werden. Den hier soll niemand manuell eingreifen.

Ist der Performance Unterschied zwischen CDM 8.5.1.x signifikant zum LTS 8.5.0.5? Da so ein SAN-System in der Regel nicht so häufig gewartet wird (FW-Updates), denke ich LTS könnte das bessere für uns sein.

Ich melde mich hier mit den Ergebnissen nach FW-Update und der HyperSwap Config.

PS: Woher war meine Muttersprache ersichtlich? 😊

"I think that is a generic message as the system is not aware the volumes are mirrored to a pool hosted on a separate system."

It is a pity, Okay thank you!

But tried and tested there is no access to a mirrored volume when the IO Enlosure is shutdown.

So I think we have to use the hyperswap modus to create a high availability mirrored volume.

A voume with automatically change the io group. Is this right ?

Yes I wrote the same:

While shutdown one enclosure the GUI Warning message is:
"Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 2 volumes." 

But the Host is able to access both volumes ?!? Maybe the GUI Warning is wrong ?

And my question is:

Does anybody know why the GUI shows the warning "…the host will not be able to access 2 volumes." 

In my opinion, the warning should not appear. (And with captcha code even more dramatic)

Thank you.

Hello Masteen,

no we have 2 enclosures with each 2 nodes. Site1=enclosure1, Site2=enclosure2, Site3=quorum

The volume should work on both enclosures (4 nodes). So if Site1 is down (enclosure 1) the host(s) should have access to this volume (enclosure 2).

In my screenshot I shutdown enclosure1. The volumes are online in enclosure2, alias site2, alias Bunker:

Thank you. We have only 1 enclosure in 2 different rooms. Both connected with 16gbit FC. If I understand you correctly, I didn't need a hyperswap. I will change the system topology. 

Does anybody know why the GUI shows the waring "…the host will not be able to access 2 volumes." 

In my first test, the ESX Host lost the volume. After configuring a VVOL Host, the host can access the volume with ony one enclosure. I think this was a path reconnect issue.

Sorry for my many newbies questions 😊

Thank you ! Yes you are right a 4-node cluster or 2 enclosure cluster 😊

 
And yes I think the context of a "striped" volume is/was not clear.

For me is a striped volume on 2 or more disks. If I lose one disk the volume is lost:

But now I understand it is a mirrored volume, each "volume copy" will be striped in its respective pool.

By the way what is the difference between Hyperswap and Mirrored Volume ?

.... 

The confusion starts with the GUI message:

After my new knowledge from the red book, I tested the system again and create VVOL hosts. 

While shutdown one enclosure the GUI message is the same:
"Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 2 volumes." 

But now the Host is able to access both volumes. Maybe the GUI Warning is wrong ?

Hello,

we got 2 new 5035 FC Storages. We done the initial setup on the first one and add the second to the first. Now we have a two-node cluster.

We create a Pool0 with all SSDs in System1 (Raid6) and a Pool1 with all SSDs in System2.

We download and start the IP Quorum App successful (connect).

Our goal is to create a realtime/live mirrored voulmes.

I create a mirrored Volume. Copy0 to Pool0 and Copy2 to Pool1.

In the details of this volume I see the volume is striped ?!? So if I power off enclosure-1 I got the waring "Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 1 volume."

May be I misunderstand mirrored Volumes ?!?

Now we add the Remote Copy license and modify the system Topology to HyperSwap. Now we got 3 sites. 1=enclosure1, 2=enclosure2, 3=quorum

I create a HyperSwap Volume in Pool0 and Pool1. In the details of the volume, I found again "Virtualization type: Striped".

And same behavior, If I shutdown enclosure1 I got the message "Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 1 volume."

If I shutdown I enclosure2 I got the warning "Powering off the control enclosure will impact system performance." And the volume is online.

But we want the volume to remain accessible on both enclosure.

Please help !

In my test environment unfortunately no. I must set the copy-volume manual online.

Of course, you can script it. But now VMWare show a new Datastore with a new ID.

This datastore (volume) have the same content, but it is a new volume.

Therefore, have to use hyperswap.

Finale Rückmeldung:

Es funktioniert mit Hyperswap 😊

Das Volume ist nun auf beiden IO Gruppen zugänglich:

Das Enclosure1 (mit Node1+2) als auch das Enclosure2 (mit Node 3+4) kann problemlos heruntergefahren oder im laufenden Betrieb getrennt werden.

Die aktiv arbeitende VM auf dem Volume bekommt davon nichts mit. Auch die Reconnects nach dem booten der Nodes funktioniert schnell und zuverlässig (Round Robin).

Leider gibt es immer noch die blöde Meldung in der IBM GUI, dass das Volume ohne Enclosure-2 nicht mehr verfügbar sein soll. Das ist aber nachweislich falsch:

Ich hoffe das hier IBM später man nachbessert.

Mit je 6x 1,92TB SAS SSD (Raid6) erreichen die beiden 5035 übrigens über 60.000 IOPS:

32 Outstanding I/Os; Default Access (100% Random, 67% Read, 2K), IBM LTS 8.5.0.5

------------------------------------------------------------------------------------------------------------------

Final feedback:

It works with Hyperswap 😊

The volume is now accessible on both IO groups:

Enclosure1 (with Node1+2) as well as Enclosure2 (with Node 3+4) can be powered down or hot disconnected. A working VM on the volume does not stop.

Unfortunately there is still the stupid message in the IBM GUI:
But this is wrong.


I hope that IBM will correct this later.

The system with 6x 1.92TB SAS SSD (Raid6) in each enclousre, reached over 60,000 IOPS:

32 outstanding I/Os; Default Access (100% Random, 67% Read, 2K), IBM LTS 8.5.0.5

Hallo Patrik,

vielen vielen Dank für die tolle Erklärung !!! 😊

Dann ist es klar, HyperSwap muss für unseren HA-Betrieb genutzt werden. Den hier soll niemand manuell eingreifen.

Ist der Performance Unterschied zwischen CDM 8.5.1.x signifikant zum LTS 8.5.0.5? Da so ein SAN-System in der Regel nicht so häufig gewartet wird (FW-Updates), denke ich LTS könnte das bessere für uns sein.

Ich melde mich hier mit den Ergebnissen nach FW-Update und der HyperSwap Config.

PS: Woher war meine Muttersprache ersichtlich? 😊

"I think that is a generic message as the system is not aware the volumes are mirrored to a pool hosted on a separate system."

It is a pity, Okay thank you!

But tried and tested there is no access to a mirrored volume when the IO Enlosure is shutdown.

So I think we have to use the hyperswap modus to create a high availability mirrored volume.

A voume with automatically change the io group. Is this right ?

Yes I wrote the same:

While shutdown one enclosure the GUI Warning message is:
"Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 2 volumes." 

But the Host is able to access both volumes ?!? Maybe the GUI Warning is wrong ?

And my question is:

Does anybody know why the GUI shows the warning "…the host will not be able to access 2 volumes." 

In my opinion, the warning should not appear. (And with captcha code even more dramatic)

Thank you.

Hello Masteen,

no we have 2 enclosures with each 2 nodes. Site1=enclosure1, Site2=enclosure2, Site3=quorum

The volume should work on both enclosures (4 nodes). So if Site1 is down (enclosure 1) the host(s) should have access to this volume (enclosure 2).

In my screenshot I shutdown enclosure1. The volumes are online in enclosure2, alias site2, alias Bunker:

Thank you. We have only 1 enclosure in 2 different rooms. Both connected with 16gbit FC. If I understand you correctly, I didn't need a hyperswap. I will change the system topology. 

Does anybody know why the GUI shows the waring "…the host will not be able to access 2 volumes." 

In my first test, the ESX Host lost the volume. After configuring a VVOL Host, the host can access the volume with ony one enclosure. I think this was a path reconnect issue.

Sorry for my many newbies questions 😊

Thank you ! Yes you are right a 4-node cluster or 2 enclosure cluster 😊

 
And yes I think the context of a "striped" volume is/was not clear.

For me is a striped volume on 2 or more disks. If I lose one disk the volume is lost:

But now I understand it is a mirrored volume, each "volume copy" will be striped in its respective pool.

By the way what is the difference between Hyperswap and Mirrored Volume ?

.... 

The confusion starts with the GUI message:

After my new knowledge from the red book, I tested the system again and create VVOL hosts. 

While shutdown one enclosure the GUI message is the same:
"Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 2 volumes." 

But now the Host is able to access both volumes. Maybe the GUI Warning is wrong ?

Hello,

we got 2 new 5035 FC Storages. We done the initial setup on the first one and add the second to the first. Now we have a two-node cluster.

We create a Pool0 with all SSDs in System1 (Raid6) and a Pool1 with all SSDs in System2.

We download and start the IP Quorum App successful (connect).

Our goal is to create a realtime/live mirrored voulmes.

I create a mirrored Volume. Copy0 to Pool0 and Copy2 to Pool1.

In the details of this volume I see the volume is striped ?!? So if I power off enclosure-1 I got the waring "Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 1 volume."

May be I misunderstand mirrored Volumes ?!?

Now we add the Remote Copy license and modify the system Topology to HyperSwap. Now we got 3 sites. 1=enclosure1, 2=enclosure2, 3=quorum

I create a HyperSwap Volume in Pool0 and Pool1. In the details of the volume, I found again "Virtualization type: Striped".

And same behavior, If I shutdown enclosure1 I got the message "Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 1 volume."

If I shutdown I enclosure2 I got the warning "Powering off the control enclosure will impact system performance." And the volume is online.

But we want the volume to remain accessible on both enclosure.

Please help !

Hello Hans,

unfortunately the weekend was too short for further tests. The system is running fine in hyperswap mode 🙂

The system passed all failure tests. Without any manual admin intervention.

I'm sure we need hyperswap. I found this IBM slide, here you see the info "same LUN ID with hyperswap":

https://docplayer.org/74947892-Vergleich-ibm-storwize-local-hyperswap-vs-ibm-svc-stretched-cluster.html

Regrads,
Thomas

In my test environment unfortunately no. I must set the copy-volume manual online.

Of course, you can script it. But now VMWare show a new Datastore with a new ID.

This datastore (volume) have the same content, but it is a new volume.

Therefore, have to use hyperswap.

Finale Rückmeldung:

Es funktioniert mit Hyperswap 😊

Das Volume ist nun auf beiden IO Gruppen zugänglich:

Das Enclosure1 (mit Node1+2) als auch das Enclosure2 (mit Node 3+4) kann problemlos heruntergefahren oder im laufenden Betrieb getrennt werden.

Die aktiv arbeitende VM auf dem Volume bekommt davon nichts mit. Auch die Reconnects nach dem booten der Nodes funktioniert schnell und zuverlässig (Round Robin).

Leider gibt es immer noch die blöde Meldung in der IBM GUI, dass das Volume ohne Enclosure-2 nicht mehr verfügbar sein soll. Das ist aber nachweislich falsch:

Ich hoffe das hier IBM später man nachbessert.

Mit je 6x 1,92TB SAS SSD (Raid6) erreichen die beiden 5035 übrigens über 60.000 IOPS:

32 Outstanding I/Os; Default Access (100% Random, 67% Read, 2K), IBM LTS 8.5.0.5

------------------------------------------------------------------------------------------------------------------

Final feedback:

It works with Hyperswap 😊

The volume is now accessible on both IO groups:

Enclosure1 (with Node1+2) as well as Enclosure2 (with Node 3+4) can be powered down or hot disconnected. A working VM on the volume does not stop.

Unfortunately there is still the stupid message in the IBM GUI:
But this is wrong.


I hope that IBM will correct this later.

The system with 6x 1.92TB SAS SSD (Raid6) in each enclousre, reached over 60,000 IOPS:

32 outstanding I/Os; Default Access (100% Random, 67% Read, 2K), IBM LTS 8.5.0.5

Hallo Patrik,

vielen vielen Dank für die tolle Erklärung !!! 😊

Dann ist es klar, HyperSwap muss für unseren HA-Betrieb genutzt werden. Den hier soll niemand manuell eingreifen.

Ist der Performance Unterschied zwischen CDM 8.5.1.x signifikant zum LTS 8.5.0.5? Da so ein SAN-System in der Regel nicht so häufig gewartet wird (FW-Updates), denke ich LTS könnte das bessere für uns sein.

Ich melde mich hier mit den Ergebnissen nach FW-Update und der HyperSwap Config.

PS: Woher war meine Muttersprache ersichtlich? 😊

"I think that is a generic message as the system is not aware the volumes are mirrored to a pool hosted on a separate system."

It is a pity, Okay thank you!

But tried and tested there is no access to a mirrored volume when the IO Enlosure is shutdown.

So I think we have to use the hyperswap modus to create a high availability mirrored volume.

A voume with automatically change the io group. Is this right ?

Yes I wrote the same:

While shutdown one enclosure the GUI Warning message is:
"Powering off the configuration node temporarily logs you out of the management GUI and the host will not be able to access 2 volumes." 

But the Host is able to access both volumes ?!? Maybe the GUI Warning is wrong ?

And my question is:

Does anybody know why the GUI shows the warning "…the host will not be able to access 2 volumes." 

In my opinion, the warning should not appear. (And with captcha code even more dramatic)

Thank you.

Hello Masteen,

no we have 2 enclosures with each 2 nodes. Site1=enclosure1, Site2=enclosure2, Site3=quorum

The volume should work on both enclosures (4 nodes). So if Site1 is down (enclosure 1) the host(s) should have access to this volume (enclosure 2).

In my screenshot I shutdown enclosure1. The volumes are online in enclosure2, alias site2, alias Bunker:

Thank you. We have only 1 enclosure in 2 different rooms. Both connected with 16gbit FC. If I understand you correctly, I didn't need a hyperswap. I will change the system topology. 

Does anybody know why the GUI shows the waring "…the host will not be able to access 2 volumes." 

In my first test, the ESX Host lost the volume. After configuring a VVOL Host, the host can access the volume with ony one enclosure. I think this was a path reconnect issue.

Sorry for my many newbies questions 😊

Thank you ! Yes you are right a 4-node cluster or 2 enclosure cluster 😊

 
And yes I think the context of a "striped" volume is/was not clear.