Update Nimble OS for Arrays that are in Synchronous Replication

[J-Philippe]

Hello to all,

do I need to pause the replication when updating Nimble OS on Arrays within a group ?

Replication is one direction : A => B

As the first updated Nimble will be in a NOS version higher than the second I'm wondering if the replication will continue to work or if I have to pause it during the update of both arrays.

Regards,

Jean-Philippe.

[Mahesh202 · Accepted solution]

Hi J-Philippe

When updating Nimble OS on arrays within a replication group, it is generally recommended to pause the replication before updating any of the arrays. This is to ensure that the replication process is not affected by any changes made to the arrays during the update.

In your scenario where replication is one direction (A => B), it is still recommended to pause replication before updating either array. The reason for this is that when you update the first Nimble array (A), it will have a higher NOS version than the second array (B), which may cause compatibility issues with the replication process.

By pausing replication before updating either array, you can ensure that the replication process is not affected by the update and that both arrays are running the same version of Nimble OS once replication is resumed.

It is also recommended to verify the replication status and health after resuming replication to ensure that the replication process is working as expected.

Hope this helps.!!

Regards
Mahesh.

[Moderator edit: removed the commands]

[Agent4046]

To answer the original question: 

- There's no easy way to "pause" synchonous replication on Nimble arrays.
- You can update the arrays while synchronous replication is running, the replication might go out of sync and will resync after communication is re-established.
Source of knowledge: a decade working as a Nimble Support Engineer.

Hello Mahesh:
There is no "replication" or "ncli" command in NimbleOS, I do not know where you got your info from.
Unfortunately the reply seems to be sourced from chatGPT and does not refer to any command supported on the nimble array.
"replication" is not a supported command on the Nimble CLI.

The way to "pause" synchronous replication is a bit more convoluted and involves going into every volume collection and removing the replication part of each schedule. It is a pain, but that's the way to do it.
You can script it using the CLI with the volcoll command:

volcoll --list to get a list of volcolls
volcoll --info <volcoll_name> | grep -i schedule --> to get the schedule names
volcoll --stop_repl <nameofvolcoll> --schedule <name_of_schedule>  --> this will stop replication for each individual schedule (settings will be lost)

This can be also done on the GUI setting replication to "none"-


As for the commands that you put in.:

Nimble OS $
Nimble OS $ replication --help
bash: replication: command not found

Nimble OS $ help
folder prottmpl useradmin
alarm group reboot userauth
alert halt route userpolicy
array help setup usersession
auditlog initiatorgrp shelf vcenter
cert ip snap version
chapuser migration snapcoll vm
ctrlr netconfig software vol
date nic sshkey volcoll
disk partner stats witness
encryptkey pe subnet
failover perfpolicy support
fc pool timezone