LVM ran out of metadata space - need help

[Red-Bull]

Hello,

i wanted to update one of my VMs and noticed i/o errors - so i checked the proxmox host and saw dmesg output:

[2797106.226966] device-mapper: space map metadata: unable to allocate new metadata block
[2797106.227006] device-mapper: thin: 253:4: metadata operation 'dm_thin_insert_block' failed: error = -28
[2797106.227035] device-mapper: thin: 253:4: aborting current metadata transaction
[2797106.231016] device-mapper: thin: 253:4: switching pool to read-only mode


lvs shows:



how can i fix this? VMs do not start anymore !

 

[Red-Bull]

i tried to delete swap space and extend metadata - but does not work ?"

 

[mailinglists]

Hi.

The first thing you can do right now is try to backup VMs from inside them using rsync or something like that.
Do not shut down or reboot them, before you do that. I'm afraid you have experienced loss of data due to bug in PM, for which they do not seem to care. See here:
https://bugzilla.proxmox.com/show_bug.cgi?id=1241

And here:
https://forum.proxmox.com/threads/is-default-install-lvm2-thin-pool-metadata-size-appropriate.31627/

Last post has a suggestion for you. Good luck!

 

[Red-Bull]

Hi,

thanks for your reply.

The problem was that i did a "yum update" during this occurs. I had no luck... i had to reboot the Proxmox host, shring the SWAP LVM space and use 1GB for metadata space. I restored the VM from backup the night before.

Really not a nice thing... i hope proxmox team take care of "meta data" size in the summary or so ?