LVM snapshot lockups on several servers since upgrading to Proxmox 2.3

[gkovacs]

Good to hear. We are also running kernel -19 over ext4 (Intel CPU, Adaptec controller) and snapshot backups work fine.

However, we have problems with quota on PVE 2.3 over ext4, please see this thread:
http://forum.proxmox.com/threads/13465-no-quota-engine-running

Do you have quota running on these servers? Is it ok for your OpenVZ guests?

 

[HostVPS]

Good to hear. We are also running kernel -19 over ext4 (Intel CPU, Adaptec controller) and snapshot backups work fine.

However, we have problems with quota on PVE 2.3 over ext4, please see this thread:
http://forum.proxmox.com/threads/13465-no-quota-engine-running

Do you have quota running on these servers? Is it ok for your OpenVZ guests?

Sorry, wish I could help, but we're not running quota on these nodes.

Just looking at your thread and I agree with your comments regarding ext4 vs ext3 - ext3 is just too slow for fscks on the large arrays we have these days, hope you get this sorted.

 

[gkovacs]

Sorry, wish I could help, but we're not running quota on these nodes.

Just looking at your thread and I agree with your comments regarding ext4 vs ext3 - ext3 is just too slow for fscks on the large arrays we have these days, hope you get this sorted.

May I ask then why are you running on ext4? The biggest reason for us was the sloooooow quota calculation after migrations and unscheduled reboots (also fsck).
If you are not running quota, what's ext4 doing for you?

Also, have you disabled quota (Resource Group Manager daemon) manually from the web interface? Because for us, it wouldn't even start.

 

[HostVPS]

May I ask then why are you running on ext4? The biggest reason for us was the sloooooow quota calculation after migrations and unscheduled reboots (also fsck).
If you are not running quota, what's ext4 doing for you?

Also, have you disabled quota (Resource Group Manager daemon) manually from the web interface? Because for us, it wouldn't even start.

As I mentioned above, fast fsck times on large RAID arrays when using ext4 was the biggest factor for me.

Have you seen https://bugzilla.openvz.org/show_bug.cgi?id=2085 ? Looks like your bug... no?

 

[gkovacs]

Have you seen https://bugzilla.openvz.org/show_bug.cgi?id=2085 ? Looks like your bug... no?

Yes it does look like my bug. Maybe even connected to yours.

Thanks for pointing it out.