Davide,
in addition to the great advise given above, I'd like to add few more details:
- The process of removing an I/O group, say a control enclosure from a clustered system, is documented in IBM Docs article "Procedure: Removing a control enclosure from a system"
- Once you have completed the steps to remove the control enclosure and changed the topology to Standard, you also should remove the site properties of the nodes in SITE1
- Last not least, if using an IP Quorum, it will need to be redeployed. With only one I/O group however, it is not needed any longer anyway.
Also the hint @Patrik Groß made with regard to "delete snapshots on site 2" indeed is an important step to consider, otherwise this could be a roadblock when it comes to removing SITE2 nodes. I suppose the false error message Patrik is referring to, is something like vdisks depending on that node. While this is not very meaningful in the first place, it is correct nevertheless, since FlashCopy maps initially created for SITE2 vdisks still will have their bitmap being maintained in that I/O group, even if the vdisk has been moved to the other I/O group.
last not least you should remove the site properties of the SITE
------------------------------
Christian Schroeder
IBM SpecV Storage Support with Passion
------------------------------
Original Message:
Sent: Thu February 23, 2023 06:41 AM
From: Davide Galbussera
Subject: Changing topology from HyperSwap to Standard
Hello Patrick, yep i've missed the lasts steps, now with yours and Nezih's Hints, that should be a good point to start, correct me if i'm wrong:
- Check the multipath of the hosts is configured following the best practices
- Check the zoning to be sure that all host are visible to v7000_SITE1
- Change all hosts site to v7000_SITE1
- Check if any flashcopies is present in the v7000_SITE2, if present, remove it.
- Change the quorum setting to preferred/winner to site 1
- Run the command "rmvolumecopy -site v7000_SITE2 volume" for all hyperswap volume
- Run the command "chsystem -topology standard"
- Run the command "rmnodecanister v7000_SITE2"
- Rescan Host Multipath
- Adjust the port configuration of the v7000_SITE1 based on your needs
------------------------------
Davide Galbussera
Original Message:
Sent: Thu February 23, 2023 02:40 AM
From: Patrik Groß
Subject: Changing topology from HyperSwap to Standard
Hello Davide,
this is a good roadmap, in addition to this you should delete snapshots on site 2 in advance, depending on the firmware these can cause a delete of the V7 on site 2 to be prohibited with a false error message. Do slowly between the steps and check everything, then it will work. The last steps would be to make sure that the quorum authority is on site 1, kick the V7 in site 2 out of the cluster, possibly adjust the port configuration of the V7 in site 1, clear the messages.
How to remove the second storage:
https://www.ibm.com/docs/en/flashsystem-7x00/7.8.x?topic=commands-leavecluster
or
https://www.ibm.com/docs/en/flashsystem-7x00/7.8.x?topic=commands-rmnodecanister
After that you should reset the V7 from site 2. Format the data media there, depending on the firmware this can also be done via the CLI.
Best regards
------------------------------
Patrik Groß
Original Message:
Sent: Wed February 22, 2023 06:59 AM
From: Davide Galbussera
Subject: Changing topology from HyperSwap to Standard
Hello, for migration purposes we need to shutdown completely a v7000 forming an HyperSwap with another v7000, for convenience I'll refer to them as v7000_SITE1 and v7000_SITE2.
These are the steps that I think i'll need to do:
- check the zoning to be sure that all host are visible to v7000_SITE1
- change all hosts site to v7000_SITE1
- remove the copy1 off all hyperswap volumes that resides on v7000_SITE2: rmvolumecopy -site v7000_SITE2 volume1
- run the command "chsystem -topology standard"
I want to be sure is that all the volumes remains online serving luns to the hosts of v7000_SITE1, there are some restriction that i need to keep in mind?
thanks
------------------------------
Davide Galbussera
------------------------------