I have noticed that every single VM is impacted. On Fri, Feb 17, 2023 at 3:14 PM Eugen Block <eblock@nde.ag> wrote:
Well, it’s the other way around: the compute nodes are the ones acquiring the locks as clients. If ceph goes down they can’t do anything with the locks until the cluster is reachable again, and sometimes a service restart is required, or a manual intervention as in this case. These things happen, the only thing that would help would probably be a stretched (or geo-redundant) ceph cluster to avoid a total failure so the cloud keeps working if one site goes down. Do you see the same impact on that many VMs or only on some of them? Or what does the last question refer to?
Zitat von Satish Patel <satish.txt@gmail.com>:
Hi Eugen,
I have a few questions before we close this thread.
- Is it normal that ceph locks images during power failure or disaster? - Shouldn't ceph should release locks automatically when VMs shutdown? - Is this a bug or natural behavior of ceph? I am worried what if i have 100s of VMs and remove lock of all of them
On Fri, Feb 17, 2023 at 10:28 AM Satish Patel <satish.txt@gmail.com> wrote:
Hi Eugen,
You saved my life!!!!!! all my vms up without any filesystem error :)
This is the correct command to remove the lock.
$ rbd lock rm -p vms ec6044e6-2231-4906-9e30-1e2e72573e64_disk "auto 139643345791728" client.1211875
On Fri, Feb 17, 2023 at 10:06 AM Satish Patel <satish.txt@gmail.com> wrote:
Hi Eugen,
I am playing with less important machine and i did following
I shutdown VM but still down following lock
root@ceph1:~# rbd lock list --image ec6044e6-2231-4906-9e30-1e2e72573e64_disk -p vms There is 1 exclusive lock on this image. Locker ID Address client.1211875 auto 139643345791728 192.168.3.12:0/2259335316
root@ceph1:~# ceph osd blacklist add 192.168.3.12:0/2259335316 blocklisting 192.168.3.12:0/2259335316 until 2023-02-17T16:00:59.399775+0000 (3600 sec)
Still I can see it in the following lock list. Am I missing something?
root@ceph1:~# rbd lock list --image ec6044e6-2231-4906-9e30-1e2e72573e64_disk -p vms There is 1 exclusive lock on this image. Locker ID Address client.1211875 auto 139643345791728 192.168.3.12:0/2259335316
On Fri, Feb 17, 2023 at 2:39 AM Eugen Block <eblock@nde.ag> wrote:
The lock is aquired automatically, you don't need to create one. I'm curious why you have that many blacklist entries, maybe that is indeed the issue here (locks are not removed). I would shutdown the corrupted VM and see if the compute node still has a lock on that image, because after shutdown it should remove the lock (automatically). If there's still a watcher or lock on that image after shutdown (rbd status vms/55dbf40b-0a6a-4bab-b3a5-b4bb74e963af_disk) you can try to blacklist the client with:
# ceph osd blacklist add client.<ID>
Then check the status again, if no watchers are present, boot the VM.
Zitat von Satish Patel <satish.txt@gmail.com>:
Hi Eugen,
This is what I did, let me know if I missed anything.
root@ceph1:~# ceph osd blacklist ls 192.168.3.12:0/0 2023-02-17T04:48:54.381763+0000 192.168.3.22:0/753370860 2023-02-17T04:47:08.185434+0000 192.168.3.22:0/2833179066 2023-02-17T04:47:08.185434+0000 192.168.3.22:0/1812968936 2023-02-17T04:47:08.185434+0000 192.168.3.22:6824/2057987683 2023-02-17T04:47:08.185434+0000 192.168.3.21:0/2756666482 2023-02-17T05:16:23.939511+0000 192.168.3.21:0/1646520197 2023-02-17T05:16:23.939511+0000 192.168.3.22:6825/2057987683 2023-02-17T04:47:08.185434+0000 192.168.3.21:0/526748613 2023-02-17T05:16:23.939511+0000 192.168.3.21:6815/2454821797 2023-02-17T05:16:23.939511+0000 192.168.3.22:0/288537807 2023-02-17T04:47:08.185434+0000 192.168.3.21:0/4161448504 2023-02-17T05:16:23.939511+0000 192.168.3.21:6824/2454821797 2023-02-17T05:16:23.939511+0000 listed 13 entries
root@ceph1:~# rbd lock list --image 55dbf40b-0a6a-4bab-b3a5-b4bb74e963af_disk -p vms There is 1 exclusive lock on this image. Locker ID Address client.268212 auto 139971105131968 192.168.3.12:0/1649312807
root@ceph1:~# ceph osd blacklist rm 192.168.3.12:0/1649312807 192.168.3.12:0/1649312807 isn't blocklisted
How do I create a lock?
On Thu, Feb 16, 2023 at 10:45 AM Eugen Block <eblock@nde.ag> wrote:
> In addition to Sean's response, this has been asked multiple times, > e.g. [1]. You could check if your hypervisors gave up the lock on the > RBDs or if they are still locked (rbd status <pool>/<image>), in that > case you might need to blacklist the clients and see if that resolves > anything. Do you have regular snapshots (or backups) to be able to > rollback in case of a curruption? > > [1] https://www.spinics.net/lists/ceph-users/msg45937.html > > > Zitat von Sean Mooney <smooney@redhat.com>: > > > On Thu, 2023-02-16 at 09:56 -0500, Satish Patel wrote: > >> Folks, > >> > >> I am running a small 3 node compute/controller with 3 node ceph storage > in > >> my lab. Yesterday, because of a power outage all my nodes went down. > After > >> reboot of all nodes ceph seems to show good health and no error (in ceph > >> -s). > >> > >> When I started using the existing VM I noticed the following errors. > Seems > >> like data loss. This is a lab machine and has zero activity on vms but > >> still loses data and the file system corrupt. Is this normal ? > > if the vm/cluster hard crashes due to the power cut yes it can. > > personally i have hit this more often with XFS then ext4 but i have > > seen it with both. > >> > >> I am not using eraser coding, does that help in this matter? > >> > >> blk_update_request: I/O error, dev sda, sector 233000 op 0x1: (WRITE) > flags > >> 0x800 phys_seg 8 prio class 0 > > > > you will proably need to rescue the isntance and repair the > > filesystem of each vm with fsck > > or similar. so boot with recue image -> repair filestem -> unrescue > > -> hardreboot/start vm if needed > > > > you might be able to mitigate this somewhat by disableing disk > > cacheing at teh qemu level but > > that will reduce performance. ceph recommenes that you use > > virtio-scis fo the device model and > > writeback cach mode. we generally recommend that too however you can > > use the disk_cachemodes option to > > chage that. > > >
https://docs.openstack.org/nova/latest/configuration/config.html#libvirt.dis...
> > > > [libvirt] > > disk_cachemodes=file=none,block=none,network=none > > > > this curreption may also have happend on the cecph cluter side. > > they have some options that can help prevent that via journaling wirtes > > > > if you can afford it i would get even a small UPS to allow a > > graceful shutdown if you have future powercuts > > to aovid dataloss issues. > > > > >