[PVE-User] Failed to start Journal Service.

Elias pve at whitenite.de
Thu Feb 15 15:56:57 CET 2018


Hello,

this problem is appearing again. I have some logs this time with a
kernel crash (not yet sure if this is linked to the IO-load problem).

My "local"-disk is a lvm-based raid1 and the "data"-lvm with vm-images
and lxc-dirs is a lvm-raid1, too. My filesystems are ext3 (local) and
xfs (data).

Please find a bunch of different log files [0] that I collected shortly
after the problem occured again. Unfortunately, the dmesg output was
already full of Journald-messages when I looked at it.
I suspect one of my LXC containers of triggering this problem but have
no evidence so far, this is only speculation.

The IO-Load is constantly around ~25% and the load is rising linear. SSH
and HTTPS are hung. The only possibility to regain access is to reboot
the server.

[0] https://pastebin.com/5GzhfCyJ


On 01/29/2018 01:05 PM, Elias wrote:
> dmesg was full of said logs. Logins and package updates didn't work either.
> The load-average rose in a linear fashion to ~1700, the IO-load was at a
> constant ~17 before I rebooted.
> Again, this did not affect VM performance.
> 
> Since the update to pve-manager/5.1-43/bdb08029 this issue did not
> appear again.
> 
> 
> On 01/24/2018 07:32 AM, Thomas Lamprecht wrote:

-- 
PGP: 5282 38E2 944B E787 8677 90D0 EB70 440A 3B1C 6C10

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 833 bytes
Desc: OpenPGP digital signature
URL: <http://pve.proxmox.com/pipermail/pve-user/attachments/20180215/96564d90/attachment.sig>


More information about the pve-user mailing list