Hi,

We have a remote site where we had recently new Flashsystem 7300's installed. The only method for the local support team to set the IP was to use a screen/keyboard in the back, and using CLI to run the satask mkcluster command. Unfortunately the IP address we gave them was wrong and we need to now reset the IP to the correct address. I'm a little confused with the different commands, both satask, and native commands, what they need to do to change the cluster MGMT IP. Can someone please advise the best CLI option(s) to reset the cluster IP to another IP address?

Cheers,

Andy

------------------------------
Andy Heath
------------------------------

Hi Andy, 

in fact you will need both satask and native (svctask) commands.
The good message is,  when you ssh into the service IP address of the current config node,
you can issue the command to change the actual cluster management IP address.
Let's find out, which is the current config node using CLI command sainfo lsservicestatus <panel_id>.
For your convenience I have combined the needed commands in this handy one-liner:

sainfo lsservicenodes -nohdr|while read panelid rest;do sainfo lsservicestatus $panelid|grep -E '^panel_name|^service_IP_address|^config'; done

The resulting output should look similar to this example:

panel_name 01-1
config_node Yes <<<<<< this is the one we are looking for >>>>>
service_IP_address 10.0.0.11
service_IP_address_6
panel_name 01-2
config_node No
service_IP_address 10.0.0.12
service_IP_address_6

Now, ssh into the service_IP_address of that node canister, which is the config node.
Then change the cluster_ip to the correct one, e.g. 10.0.0.10 like this:

chsystemip -clusterip 10.0.0.10 -gw 10.0.0.254 -mask 255.0.0.0 -port 1

Best regards, 
Christian Schroeder

------------------------------
Christian Schroeder
IBM SpecV Storage Support with Passion
------------------------------

Original Message:
Sent: Tue April 04, 2023 10:17 AM
From: Andy Heath
Subject: New Flashsystem satask mkcluster IP error

Hi,

We have a remote site where we had recently new Flashsystem 7300's installed. The only method for the local support team to set the IP was to use a screen/keyboard in the back, and using CLI to run the satask mkcluster command. Unfortunately the IP address we gave them was wrong and we need to now reset the IP to the correct address. I'm a little confused with the different commands, both satask, and native commands, what they need to do to change the cluster MGMT IP. Can someone please advise the best CLI option(s) to reset the cluster IP to another IP address?

Cheers,

Andy

------------------------------
Andy Heath
------------------------------

Hello Andy,

Some context for you:

sainfo and satask commands are the patterns used for the service cli.  This is a special CLI that exists on every node/node canister and is used for:
1: Initial configuration of the node
2: Maintenance or service operations on the node that must be accessible when the system is not active in a cluster
3: specific node ip operations that are not available in the clustered system cli

The service  CLI is only accessible to the superuser (the local user that is default on every node and is available outside of a cluster with maximum privilege) and is generally accessed on the Service IP (although you can access it on the technician port or management port, it is a service that runs on every node so is available on all management configured interfaces).  If you have not set a service IP, I highly recommend you do so for each node/canister in your cluster.  there is only 1 service IP per node/nodecanister.  There is also a service GUI which makes use of the service CLI and can be used to adjust the settings or perform the operations per node for service of the individual node, or a partner node on the SAN/PCIe network (these are known as remote service CLIs and you basically just provide the target nodes panel_name to the GUI/service CLI command line).

The other CLI commands (previously svctask/svcinfo but now with no prefix usually of the form ls$COMMAND mk$COMMAND rmCOMMAND ch$COMMAND) are known as the cluster CLI or system CLI, this is the primary command line management interface, it is what is mirrored in the RESTAPI and is used by the GUI to do day to day operational tasks, monitoring and most directed maintenance procedures.

This CLI is hosted by the current configuration node on the System or Cluster IP address, and if configured the secondary system ip on Ethernet ports 1 and 2 (if configured). There is only 1 system IP and 1 secondary system IP that can be configured per system (so 2-8 node or node canisters depending on the Virtualize system in question will share 1 primary system IP which will host the primary management CLI).  The management GUI uses this primary or main CLI to perform its tasks/operations.  The RESTAPI is derived from the main/system CLI as previously stated.  This will move between nodes in the system if a node failover occurs.


Having said ALL of the above:

If you want to fix your IPs, you will to have a  person on site connect to the technician port, and go to the GUI to change this (alternatively if you can access the IP via a host or system on site you can adjust the ip settings using the one of the GUIs or the CLI).  The system IP can be changed both from the service CLI/GUI and the main CLI/GUI.  The service IP can only be changed from the service GUI/service CLI.


Internally they are actually two different exectables with different functions/responsibilities.  The Service operations as stated before should be very limited to initial set up and then service/data collect for nodes in service/not in an active cluster.


Some references that might help:
https://www.ibm.com/docs/en/flashsystem-9x00/8.5.x?topic=t-user-interfaces-servicing-your-system-4
https://www.ibm.com/docs/en/flashsystem-9x00/8.5.x?topic=cli-using-cli-4

https://www.ibm.com/docs/en/flashsystem-9x00/8.5.x?topic=uisys-command-line-interface-4
https://www.ibm.com/docs/en/flashsystem-9x00/8.5.x?topic=uisys-service-assistant-interface-4

Please note that the CLI is the same general code across all Storage Virtualize products, with minor difference to account for hardware (node in SVC versus nodecanister in FlashSystem and cloudy things in SV4PC).

I hope that helps!

------------------------------
Evelyn Perez
------------------------------

Original Message:
Sent: Tue April 04, 2023 10:17 AM
From: Andy Heath
Subject: New Flashsystem satask mkcluster IP error

Hi,

We have a remote site where we had recently new Flashsystem 7300's installed. The only method for the local support team to set the IP was to use a screen/keyboard in the back, and using CLI to run the satask mkcluster command. Unfortunately the IP address we gave them was wrong and we need to now reset the IP to the correct address. I'm a little confused with the different commands, both satask, and native commands, what they need to do to change the cluster MGMT IP. Can someone please advise the best CLI option(s) to reset the cluster IP to another IP address?

Cheers,

Andy

------------------------------
Andy Heath
------------------------------

I thought service addresses could also be updated from main CLI/GUI :

------------------------------
Hans Populaire
------------------------------

Original Message:
Sent: Wed April 05, 2023 05:33 AM
From: Evelyn Perez
Subject: New Flashsystem satask mkcluster IP error

Hello Andy,

Some context for you:

sainfo and satask commands are the patterns used for the service cli.  This is a special CLI that exists on every node/node canister and is used for:
1: Initial configuration of the node
2: Maintenance or service operations on the node that must be accessible when the system is not active in a cluster
3: specific node ip operations that are not available in the clustered system cli

The service  CLI is only accessible to the superuser (the local user that is default on every node and is available outside of a cluster with maximum privilege) and is generally accessed on the Service IP (although you can access it on the technician port or management port, it is a service that runs on every node so is available on all management configured interfaces).  If you have not set a service IP, I highly recommend you do so for each node/canister in your cluster.  there is only 1 service IP per node/nodecanister.  There is also a service GUI which makes use of the service CLI and can be used to adjust the settings or perform the operations per node for service of the individual node, or a partner node on the SAN/PCIe network (these are known as remote service CLIs and you basically just provide the target nodes panel_name to the GUI/service CLI command line).

The other CLI commands (previously svctask/svcinfo but now with no prefix usually of the form ls$COMMAND mk$COMMAND rmCOMMAND ch$COMMAND) are known as the cluster CLI or system CLI, this is the primary command line management interface, it is what is mirrored in the RESTAPI and is used by the GUI to do day to day operational tasks, monitoring and most directed maintenance procedures.

This CLI is hosted by the current configuration node on the System or Cluster IP address, and if configured the secondary system ip on Ethernet ports 1 and 2 (if configured). There is only 1 system IP and 1 secondary system IP that can be configured per system (so 2-8 node or node canisters depending on the Virtualize system in question will share 1 primary system IP which will host the primary management CLI).  The management GUI uses this primary or main CLI to perform its tasks/operations.  The RESTAPI is derived from the main/system CLI as previously stated.  This will move between nodes in the system if a node failover occurs.


Having said ALL of the above:

If you want to fix your IPs, you will to have a  person on site connect to the technician port, and go to the GUI to change this (alternatively if you can access the IP via a host or system on site you can adjust the ip settings using the one of the GUIs or the CLI).  The system IP can be changed both from the service CLI/GUI and the main CLI/GUI.  The service IP can only be changed from the service GUI/service CLI.


Internally they are actually two different exectables with different functions/responsibilities.  The Service operations as stated before should be very limited to initial set up and then service/data collect for nodes in service/not in an active cluster.


Some references that might help:
https://www.ibm.com/docs/en/flashsystem-9x00/8.5.x?topic=t-user-interfaces-servicing-your-system-4
https://www.ibm.com/docs/en/flashsystem-9x00/8.5.x?topic=cli-using-cli-4

https://www.ibm.com/docs/en/flashsystem-9x00/8.5.x?topic=uisys-command-line-interface-4
https://www.ibm.com/docs/en/flashsystem-9x00/8.5.x?topic=uisys-service-assistant-interface-4

Please note that the CLI is the same general code across all Storage Virtualize products, with minor difference to account for hardware (node in SVC versus nodecanister in FlashSystem and cloudy things in SV4PC).

I hope that helps!

------------------------------
Evelyn Perez

Original Message:
Sent: Tue April 04, 2023 10:17 AM
From: Andy Heath
Subject: New Flashsystem satask mkcluster IP error

Hi,

We have a remote site where we had recently new Flashsystem 7300's installed. The only method for the local support team to set the IP was to use a screen/keyboard in the back, and using CLI to run the satask mkcluster command. Unfortunately the IP address we gave them was wrong and we need to now reset the IP to the correct address. I'm a little confused with the different commands, both satask, and native commands, what they need to do to change the cluster MGMT IP. Can someone please advise the best CLI option(s) to reset the cluster IP to another IP address?

Cheers,

Andy

------------------------------
Andy Heath
------------------------------