You are not logged in.
After zxg_sync was churning for a good 2 hours without doing anything else than using all CPU 100% I decided to reboot the NAS. Shutdown was halted at the end where two ZFS devices could not be unmounted and the watchdog eventually timed out.
What could be the cause of zxg_sync not responding and using all CPU? my 3T NAS is only used for about 1.7T. Earlier today I did perform quite som backup maintenance on this NAS: As far as I remember the raidz1 was in good order and now harddisk were failing.
After reboot ZFS seems to resync the devices and the ZFS daemon is starting for 30min now, causing timeout messages in the boot screen.
As the ZFS is only containing the storeage devices, the OS is on a seperate SSD, I wonder why the boot process waits for the ZFS daemon to start. It is not really needed.
EDIT: I did attempt to remove the cache from the pool just before. This might have caused a little crisis in it.
Last edited by theking2 (2014-12-24 19:11:10)
archlinux on a Gigabyte C1037UN-EU, 16GiB
a Promise PDC40718 based ZFS set
root on a Samsung SSD PB22-J
running LogitechMediaServer(-git), Samba, MiniDLNA, TOR
Offline
It might be that also that have a record size of 4k killed performance. The reason I used 4k was that this filesystem was destined to contain file history and time machine files. That would mean a lot of smallish files and a small recordsize would benefit there.
I've created a new backup devices with a record size of 32k which seems to behave much better. And I must add that the 2G RAM probably is taxed by zfs.
archlinux on a Gigabyte C1037UN-EU, 16GiB
a Promise PDC40718 based ZFS set
root on a Samsung SSD PB22-J
running LogitechMediaServer(-git), Samba, MiniDLNA, TOR
Offline