IBM FlashSystem

Hello,

we got two new FS-5300 an create a simple PBHA System (32GB Fibre Channel). 

Two Windows 2025 Hosts are directly connected to both systems. Each Win2025 Hosts with 4 32GB FC Link to each node.
(HPE DL380 Gen11, HPE StoreFabric SN1610Q HBA)

In normal operation Windows we see 4 paths to the storage:

Now we shutdown the IBM Preferred management system. The second System works well. The Windows Hosts didn't lost the disks / connections. Now Windows show only 2 two links. That is correct:

Now we boot the shutdown IBM System again. The IBM dashboard gets green. But the Windows Hosts have furthermore only 2 paths to the storage system. The 2 lost paths will not be restored (degraded).

If the second system fails now, we lose the connection to the storage.

If we reboot the hosts, we got again 4 links (redundant). We can also reconnect the lost paths by deactivating and activating the FC HBA in the device manager.

But why is there no automatic reconnect? Have we forgotten something? Different MSDSM settings do not help:

Thank you,
Thomas

Hi Thomas,

Did you check SSIC (IBM System Storage Interoperation Center) for your configuration. 

https://www.ibm.com/systems/support/storage/ssic/interoperability

Hello,

we got two new FS-5300 an create a simple PBHA System (32GB Fibre Channel). 

Two Windows 2025 Hosts are directly connected to both systems. Each Win2025 Hosts with 4 32GB FC Link to each node.
(HPE DL380 Gen11, HPE StoreFabric SN1610Q HBA)

In normal operation Windows we see 4 paths to the storage:

Now we shutdown the IBM Preferred management system. The second System works well. The Windows Hosts didn't lost the disks / connections. Now Windows show only 2 two links. That is correct:

Now we boot the shutdown IBM System again. The IBM dashboard gets green. But the Windows Hosts have furthermore only 2 paths to the storage system. The 2 lost paths will not be restored (degraded).

If the second system fails now, we lose the connection to the storage.

If we reboot the hosts, we got again 4 links (redundant). We can also reconnect the lost paths by deactivating and activating the FC HBA in the device manager.

But why is there no automatic reconnect? Have we forgotten something? Different MSDSM settings do not help:

Thank you,
Thomas

Hi Nezih,

yes, the configuration is correct. I tried again today with Windows Server 2022 instead of 2025. Unfortunately the same behavior...

FC cables can be unplugged during operation. Even if they are plugged back in after 10 minutes, there is an immediate reconnect.

But if the IBM system is not restarted there is no reconnect.

Hi Thomas,

Did you check SSIC (IBM System Storage Interoperation Center) for your configuration. 

https://www.ibm.com/systems/support/storage/ssic/interoperability

Hello,

we got two new FS-5300 an create a simple PBHA System (32GB Fibre Channel). 

Two Windows 2025 Hosts are directly connected to both systems. Each Win2025 Hosts with 4 32GB FC Link to each node.
(HPE DL380 Gen11, HPE StoreFabric SN1610Q HBA)

In normal operation Windows we see 4 paths to the storage:

Now we shutdown the IBM Preferred management system. The second System works well. The Windows Hosts didn't lost the disks / connections. Now Windows show only 2 two links. That is correct:

Now we boot the shutdown IBM System again. The IBM dashboard gets green. But the Windows Hosts have furthermore only 2 paths to the storage system. The 2 lost paths will not be restored (degraded).

If the second system fails now, we lose the connection to the storage.

If we reboot the hosts, we got again 4 links (redundant). We can also reconnect the lost paths by deactivating and activating the FC HBA in the device manager.

But why is there no automatic reconnect? Have we forgotten something? Different MSDSM settings do not help:

Thank you,
Thomas

Hi Thomas,

It is probably worth opening a support ticket with debug data so the team can take a look at what is causing this behaviour.

Thanks,

Philip

Hi Nezih,

yes, the configuration is correct. I tried again today with Windows Server 2022 instead of 2025. Unfortunately the same behavior...

FC cables can be unplugged during operation. Even if they are plugged back in after 10 minutes, there is an immediate reconnect.

But if the IBM system is not restarted there is no reconnect.

Hi Thomas,

Did you check SSIC (IBM System Storage Interoperation Center) for your configuration. 

https://www.ibm.com/systems/support/storage/ssic/interoperability

Hello,

we got two new FS-5300 an create a simple PBHA System (32GB Fibre Channel). 

Two Windows 2025 Hosts are directly connected to both systems. Each Win2025 Hosts with 4 32GB FC Link to each node.
(HPE DL380 Gen11, HPE StoreFabric SN1610Q HBA)

In normal operation Windows we see 4 paths to the storage:

Now we shutdown the IBM Preferred management system. The second System works well. The Windows Hosts didn't lost the disks / connections. Now Windows show only 2 two links. That is correct:

Now we boot the shutdown IBM System again. The IBM dashboard gets green. But the Windows Hosts have furthermore only 2 paths to the storage system. The 2 lost paths will not be restored (degraded).

If the second system fails now, we lose the connection to the storage.

If we reboot the hosts, we got again 4 links (redundant). We can also reconnect the lost paths by deactivating and activating the FC HBA in the device manager.

But why is there no automatic reconnect? Have we forgotten something? Different MSDSM settings do not help:

Thank you,
Thomas

Okay I opened a case.

Today I installed ESXi 8. No problems with reconnects. Only Windows 2022 / 2025 

Hi Thomas,

It is probably worth opening a support ticket with debug data so the team can take a look at what is causing this behaviour.

Thanks,

Philip

Hi Nezih,

yes, the configuration is correct. I tried again today with Windows Server 2022 instead of 2025. Unfortunately the same behavior...

FC cables can be unplugged during operation. Even if they are plugged back in after 10 minutes, there is an immediate reconnect.

But if the IBM system is not restarted there is no reconnect.

Hi Thomas,

Did you check SSIC (IBM System Storage Interoperation Center) for your configuration. 

https://www.ibm.com/systems/support/storage/ssic/interoperability

Hello,

we got two new FS-5300 an create a simple PBHA System (32GB Fibre Channel). 

Two Windows 2025 Hosts are directly connected to both systems. Each Win2025 Hosts with 4 32GB FC Link to each node.
(HPE DL380 Gen11, HPE StoreFabric SN1610Q HBA)

In normal operation Windows we see 4 paths to the storage:

Now we shutdown the IBM Preferred management system. The second System works well. The Windows Hosts didn't lost the disks / connections. Now Windows show only 2 two links. That is correct:

Now we boot the shutdown IBM System again. The IBM dashboard gets green. But the Windows Hosts have furthermore only 2 paths to the storage system. The 2 lost paths will not be restored (degraded).

If the second system fails now, we lose the connection to the storage.

If we reboot the hosts, we got again 4 links (redundant). We can also reconnect the lost paths by deactivating and activating the FC HBA in the device manager.

But why is there no automatic reconnect? Have we forgotten something? Different MSDSM settings do not help:

Thank you,
Thomas

What is the firmware version?

At firmware v8700

operating systems supported(8700)
Policy-based high availability supports a reduced set of the operating systems supported by IBM Storage Virtualize.

The following operating systems are supported with policy-based high availability when connected using Fibre Channel SCSI:
- Red Hat Enterprise Linux, including with KVM (all supported versions).
- SUSE Linux Enterprise (all supported versions).
- IBM AIX (version 7.2 and later)
- Linux on IBM Z (all supported versions and distributions).
- VMware ESXi (all supported versions).

At firmware v8720

operating systems supported(8720)
Policy-based high availability supports a reduced set of the operating systems that are supported by IBM Storage Virtualized.
The following operating systems are supported with policy-based high availability when connected using Fibre Channel SCSI:
- Red Hat Enterprise Linux, including with KVM (all supported versions).
- SUSE Linux Enterprise (all supported versions).
- IBM AIX (version 7.2 and later)
- Linux on IBM Z (all supported versions and distributions).
- VMware ESXi (all supported versions).
- Microsoft Windows Server, including Hyper-V (all supported versions). This requires both systems to be running IBM Storage Virtualize 8.7.2 or later

IBM Storage Virtualize 8.7.2 adds support for iSCSI with policy-based high availability. 
The following operating systems are supported with policy-based high availability when connected using iSCSI:
- Red Hat Enterprise Linux, including with KVM (all supported versions).
- SUSE Linux Enterprise (all supported versions).
- VMware ESXi (all supported versions).

Okay I opened a case.

Today I installed ESXi 8. No problems with reconnects. Only Windows 2022 / 2025 

Hi Thomas,

It is probably worth opening a support ticket with debug data so the team can take a look at what is causing this behaviour.

Thanks,

Philip

Hi Nezih,

yes, the configuration is correct. I tried again today with Windows Server 2022 instead of 2025. Unfortunately the same behavior...

FC cables can be unplugged during operation. Even if they are plugged back in after 10 minutes, there is an immediate reconnect.

But if the IBM system is not restarted there is no reconnect.

Hi Thomas,

Did you check SSIC (IBM System Storage Interoperation Center) for your configuration. 

https://www.ibm.com/systems/support/storage/ssic/interoperability

Hello,

we got two new FS-5300 an create a simple PBHA System (32GB Fibre Channel). 

Two Windows 2025 Hosts are directly connected to both systems. Each Win2025 Hosts with 4 32GB FC Link to each node.
(HPE DL380 Gen11, HPE StoreFabric SN1610Q HBA)

In normal operation Windows we see 4 paths to the storage:

Now we shutdown the IBM Preferred management system. The second System works well. The Windows Hosts didn't lost the disks / connections. Now Windows show only 2 two links. That is correct:

Now we boot the shutdown IBM System again. The IBM dashboard gets green. But the Windows Hosts have furthermore only 2 paths to the storage system. The 2 lost paths will not be restored (degraded).

If the second system fails now, we lose the connection to the storage.

If we reboot the hosts, we got again 4 links (redundant). We can also reconnect the lost paths by deactivating and activating the FC HBA in the device manager.

But why is there no automatic reconnect? Have we forgotten something? Different MSDSM settings do not help:

Thank you,
Thomas

In addition to this. You will found the supported OS for PbHA in IBM DOCS. Windows is only supported starting 8.7.2(8.7.2.1). There you will also find updates if available.

PbHA supported OS  https://www.ibm.com/docs/en/flashsystem-7x00/8.7.x?topic=availability-supported-host-operating-systems-high

What is the firmware version?

At firmware v8700

operating systems supported(8700)
Policy-based high availability supports a reduced set of the operating systems supported by IBM Storage Virtualize.

The following operating systems are supported with policy-based high availability when connected using Fibre Channel SCSI:
- Red Hat Enterprise Linux, including with KVM (all supported versions).
- SUSE Linux Enterprise (all supported versions).
- IBM AIX (version 7.2 and later)
- Linux on IBM Z (all supported versions and distributions).
- VMware ESXi (all supported versions).

At firmware v8720

operating systems supported(8720)
Policy-based high availability supports a reduced set of the operating systems that are supported by IBM Storage Virtualized.
The following operating systems are supported with policy-based high availability when connected using Fibre Channel SCSI:
- Red Hat Enterprise Linux, including with KVM (all supported versions).
- SUSE Linux Enterprise (all supported versions).
- IBM AIX (version 7.2 and later)
- Linux on IBM Z (all supported versions and distributions).
- VMware ESXi (all supported versions).
- Microsoft Windows Server, including Hyper-V (all supported versions). This requires both systems to be running IBM Storage Virtualize 8.7.2 or later

IBM Storage Virtualize 8.7.2 adds support for iSCSI with policy-based high availability. 
The following operating systems are supported with policy-based high availability when connected using iSCSI:
- Red Hat Enterprise Linux, including with KVM (all supported versions).
- SUSE Linux Enterprise (all supported versions).
- VMware ESXi (all supported versions).

Okay I opened a case.

Today I installed ESXi 8. No problems with reconnects. Only Windows 2022 / 2025 

Hi Thomas,

It is probably worth opening a support ticket with debug data so the team can take a look at what is causing this behaviour.

Thanks,

Philip

Hi Nezih,

yes, the configuration is correct. I tried again today with Windows Server 2022 instead of 2025. Unfortunately the same behavior...

FC cables can be unplugged during operation. Even if they are plugged back in after 10 minutes, there is an immediate reconnect.

But if the IBM system is not restarted there is no reconnect.

Hi Thomas,

Did you check SSIC (IBM System Storage Interoperation Center) for your configuration. 

https://www.ibm.com/systems/support/storage/ssic/interoperability

Hello,

we got two new FS-5300 an create a simple PBHA System (32GB Fibre Channel). 

Two Windows 2025 Hosts are directly connected to both systems. Each Win2025 Hosts with 4 32GB FC Link to each node.
(HPE DL380 Gen11, HPE StoreFabric SN1610Q HBA)

In normal operation Windows we see 4 paths to the storage:

Now we shutdown the IBM Preferred management system. The second System works well. The Windows Hosts didn't lost the disks / connections. Now Windows show only 2 two links. That is correct:

Now we boot the shutdown IBM System again. The IBM dashboard gets green. But the Windows Hosts have furthermore only 2 paths to the storage system. The 2 lost paths will not be restored (degraded).

If the second system fails now, we lose the connection to the storage.

If we reboot the hosts, we got again 4 links (redundant). We can also reconnect the lost paths by deactivating and activating the FC HBA in the device manager.

But why is there no automatic reconnect? Have we forgotten something? Different MSDSM settings do not help:

Thank you,
Thomas