XCP-ng 8.3 updates announcements and testing

jr-m4

Rebooting hosts and/or VMs doesn't solve the problems.
As @andrew said. This one got ugly. But this is also why we test stuff

Update:
Doing a "Rescan all disks" on the SR (lvmohba), seems to help out a little bit. But not everything.
Some VMs were able to be started again. But others are still showing problems and erroring out

"The VDI is not available [opterr=['MAP_DUPLICATE_KEY', 'VDI', 'sm_config', 'OpaqueRef:7---------2c1b', 'activating']]",

flakpyro

@jr-m4 I had this issue on one of the pools i tested on as well and was able to fix the issue by following the fix here:
https://xcp-ng.org/forum/topic/10603/sr_backend_failure_46-the-vdi-is-not-available-not-detached-cleanly/3

dthenot

@flakpyro Hello,

It would be preferred to use another method.
You can use the resetvdis.py script in the /opt/xensource/sm/ to reset values instead,
using vdi-forget is unadvised in this case, it can cause issues by losing information that are stored in XAPI about VDIs.

/opt/xensource/sm/resetvdis.py single <VDI UUID>

jr-m4

@dthenot said:

@flakpyro Hello,

It would be preferred to use another method.
You can use the resetvdis.py script in the /opt/xensource/sm/ to reset values instead,
using vdi-forget is unadvised in this case, it can cause issues by losing information that are stored in XAPI about VDIs.
/opt/xensource/sm/resetvdis.py single <VDI UUID>

I was just about to look at maybe doing what @flakpyro suggested. More likely I would've just restored from backup. As this is dev-instances, and non-important.

But your resetvdis.pysuggestion solved it entirely! Thanks for the quick tip!

jr-m4

Now trying to see if pool works with generic operations.

Trying to do a RPR on new-xcp-dev. And I'm met with

Ofcourse there are no backups being run. They are even disabled right now. And the only active task happening is "Garbage collection" on the main SR.

pool.rollingReboot
{
  "pool": "1f16dad2---------------6c8b8b7",
  "bypassBackupCheck": false,
  "shutdownPinnedVms": false
}
{
  "code": "CANNOT_EVACUATE_HOST",
  "params": [
    "VM_LACKS_FEATURE,OpaqueRef:aa5210c4-a------------1b80-9bca5d0d0f8d|OPERATION_BLOCKED,OpaqueRef:c7a08---------------a0--------------9306,VM migration is blocked during backup"
  ],
  "call": {
    "duration": 23,
    "method": "host.assert_can_evacuate",
    "params": [
      "* session id *",
      "OpaqueRef:0145019c-2a32-13bb-5951-3bca9e59e63a"
    ]
  },
  "message": "CANNOT_EVACUATE_HOST(VM_LACKS_FEATURE,OpaqueRef:aa52-----------------------9bca5d0d0f8d|OPERATION_BLOCKED,OpaqueRef:c7a0------------------------306,VM migration is blocked during backup)",
  "name": "XapiError",
  "stack": "XapiError: CANNOT_EVACUATE_HOST(VM_LACKS_FEATURE,OpaqueRef:aa521-------------------8d|OPERATION_BLOCKED,OpaqueRef:c7a08c--------------------------9306,VM migration is blocked during backup)
    at XapiError.wrap (file:///opt/xen-orchestra/packages/xen-api/_XapiError.mjs:16:12)
    at file:///opt/xen-orchestra/packages/xen-api/transports/json-rpc.mjs:38:21
    at runNextTicks (node:internal/process/task_queues:65:5)
    at processImmediate (node:internal/timers:541:9)"
}

Update:
I believe this might be related due to a VDI being orphaned, and trying to delete it results in VDI_NOT_MANAGED-error. Trying a new rescan of disks on the SR, and hopefully garbage collection will finish until Monday. Otherwise I will somehow try to forcefully destroy this VDI.

Update 2:
Managed to do a RPR now.
Seems that the flag "Prevent Migration" was applied to one of the VMs. My guess is that this is applied when backups are taken. And since the backups broke with the updates. The flag never got cleared.
I did a quick visual confirmation to see that my guess of this behaviour seems correct. And it indeed sets the "Prevent migration"-flag whilst the backup runs.

Running backups on my pools to see if functionality is working as intended again.

Update 3:
Maybe related, maybe not. But my backup-tests are piling up with lots of VBD.unplug_force, that have gone on for <=30 minutes. And this is on both pools. And it seems very much so that the backup-jobs doesn't complete properly. This could also be entirely XO-related, since there has been several commits towards backups as well.

Update 4:
Updated XO from source to the latest commit d2ffe, and things have not improved.
Pretty much, many actions with VDIs are problematic. I'll happily provide logs. But will only do so if I get somewhere to upload them to.
I'll be taking the weekend now. So won't be able to respond before monday,

JeffBerntsen

Updated and running on my test pool and twinstor test pair. I don't use XO for backups so I'm not seeing the possible issues associated with that but all else seems to be running just fine.

gduperrey

The storage and XAPI teams are currently investigating the issues that have been reported.

There is no final solution yet, but we are getting closer. Unfortunately, there is no ETA at this time.

Thank you very much for all your testing and feedback!

olivierlambert

Thanks everyone for testing! As you can all see here, it’s been extremely useful

acebmxer

This post is deleted!

lmolina

New packages in `xcp-ng-testing` for the next XCP-ng 8.3 LTS maintenance update

This is a follow-up of the the previous post and addresses the issue reported by our testers.

First of all, a huge thanks to all testers who allowed us to catch this backup issue early:
@acebmxer, @jr-m4, @bufanda, @flakpyro, @Andrew.

We decided to revert the change that improves the performance of xapi-nbd since it caused the backup issue. This change will be reworked and refined, then presented to you again in a future update, without this issue. While investigating, we also found a bug in blktap: in rare cases, the process handling a VM disk (tapdisk) could crash during a backup and leave the disk impossible to detach. The new blktap version fixes it.

Versions

blktap: 3.55.5-12.2.xcpng8.3 -> blktap-3.55.5-12.3.xcpng8.3
xapi: 26.1.21-1.1.xcpng8.3 -> 26.1.21-1.2.xcpng8.3

Test on XCP-ng 8.3

yum clean metadata --enablerepo=xcp-ng-testing,xcp-ng-candidates
yum update --enablerepo=xcp-ng-testing,xcp-ng-candidates
reboot

If you reverted the previous candidates with yum history undo, update normally with the commands above. Note that downgrading is not recommended. If you disabled NBD in your backup jobs as a workaround, please enable it again after the update.

The usual update rules apply: pool coordinator first, etc.

What to test

If you have already carried out the initial updates and performed the tests mentioned in our previous post, please confirm that the XO backups are now back to a working state.

If you did not run the initial update and have not yet performed any tests, please be sure to review the "what to test" section of the the previous post.

Test window before official release of the updates

~7 days