Hello !
I've been preparing a PVE host for several weeks and I was almost ready to put it in production. I added a 2TB disk to use for vmdata using the Administration Guide, using
thin provisioning :
That differs from the Guide with the addition of "-c 256K" since I couln't create the thin pool otherwise... But there are no indication about its ideal value.
After adding all my VMs I started a job generating A LOT of data in one of them (the VM had a 256GB virtual disk) but a few minutes after that I got a lot of errors :
Constantly repeated. As long as I try to start a VM, I get these errors.
At first I thought it was due to a hardware failure. However after both digging on the net and my system logs, looking at commands results such as smartctl, I am convinced the problem is tied to the Thin-Pool. Indeed, in the logs I could find :
So... What do you think are my options at this point ?
The Thin-Pool takes up all of the disk : how can I resize it... without loosing data ?
Maybe by growing the chunk size and having less chunks if there is a command like that ?
Note: I've just read another thread that ends with the conclusion "If you run out of space for metatada, all data will be lost. Backup your VMs as soon as possible." I only have 1 VM I'd like to backup. Could it be considered to remove them all but the one I'd like to keep, make a backup and reconfigure the thin-pool ?
Thanks in advance...
I've been preparing a PVE host for several weeks and I was almost ready to put it in production. I added a 2TB disk to use for vmdata using the Administration Guide, using
thin provisioning :
Code:
lvcreate -L 1.8T -T -c 256K -n vmstore vmdata
After adding all my VMs I started a job generating A LOT of data in one of them (the VM had a 256GB virtual disk) but a few minutes after that I got a lot of errors :
[...]
Jul 2 12:04:42 pve kernel: [ 147.745197] sd 0:0:0:0: [sda] tag#5 FAILED Result: hostbyte=DID_BAD_TARGET driverbyte=DRIVER_OK
Jul 2 12:04:42 pve kernel: [ 147.745199] sd 0:0:0:0: [sda] tag#5 CDB: Read(10) 28 00 00 00 08 00 00 01 00 00
[...]
Constantly repeated. As long as I try to start a VM, I get these errors.
At first I thought it was due to a hardware failure. However after both digging on the net and my system logs, looking at commands results such as smartctl, I am convinced the problem is tied to the Thin-Pool. Indeed, in the logs I could find :
Jul 2 12:22:31 pve lvm[472]: WARNING: Thin pool vmdata-vmstore-tpool metadata is now 100.00% full.
Jul 2 12:22:31 pve lvm[472]: WARNING: Thin pool vmdata-vmstore-tpool data is now 100.00% full.
So... What do you think are my options at this point ?
The Thin-Pool takes up all of the disk : how can I resize it... without loosing data ?
Maybe by growing the chunk size and having less chunks if there is a command like that ?
Note: I've just read another thread that ends with the conclusion "If you run out of space for metatada, all data will be lost. Backup your VMs as soon as possible." I only have 1 VM I'd like to backup. Could it be considered to remove them all but the one I'd like to keep, make a backup and reconfigure the thin-pool ?
Thanks in advance...
Last edited: