dirtycajunSo i have a zfs mount that is mounted at /mnt/media and a ext4 mount mounted at /mnt/media/storage. I need to ensure that the ext4 mount does not mount until zfs mounts. Is there a more graceful way then @reboot crontab hacky script03:49
tdsdirtycajun: what version of ubuntu and zfs?03:54
sarnolddirtycajun: if /mnt/media is mounted by zfs and /mnt/media/storage is listed in /etc/fstab, there's a pretty good chance the right thing will already happen03:54
sarnolddirtycajun: if /mnt/media is LEGACY then you ought to be able to stick both lines in /etc/fstab and I'd expect that to work too03:55
dirtycajunUbuntu 18.04.3 and zfs 0.7.5-1ubuntu16.403:56
dirtycajununfortunately it literally NEVER does it in the right order. And since the second happens first, the first cannot happen as its in use03:56
dirtycajunand yes, zfs is mounted by zfs and ext4 is mounted by fstab in this case03:57
tdsah, so before zfs-mount-generator, i think that's only in 0.8 (and actually fixes this)03:57
tdsyou could probably use the x-systemd options in fstab so your mount only happens after zfs-mount.service though?03:58
dirtycajuni have that currently tho03:59
dirtycajunand damn. does 20.04 bundle with 8.* zfs ?03:59
dirtycajunerm 0.8.*04:00
sarnoldyes, that the moment it's a 0.8.2 https://launchpad.net/ubuntu/+source/zfs-linux04:00
dirtycajunso hack it for 4 months then upgrade. could be worse04:00
dirtycajunthank yall.04:01
tdshmm, that is odd - and the generated mount unit has requires set properly and everything?04:03
sarnolddirtycajun: if you've got the time this might be worth a bug report; 18.04 will live on for a while, and maybe it'd be nice for this to be fixed (no promises of course, but an unfiled bug can't be fixed :)04:03
dirtycajunTrue. I would need to dig through bootup and logs to see exactly where its shitting the bed though to file the report. Its a longstanding issue. It actually causes castcading problems that i have fixed like docker starting which lets k8s pods schedule which cant reach a mount etc04:04
dirtycajuntds looks like 0.7.13 is the release of zfs-mount-generator04:06
dirtycajunstill not at that freeze but an fyi04:06
sarnolddirtycajun: ubuntu-bug zfsutils-linux    will probably get to the right place04:07
lordievaderGood morning07:26
tilerenderingnot sure whether this is the right channel to ask, but: am I correct in the assumption that Ryzen-based servers are not really made for that purpose? i.e. the Ryzen being a workstation cpu (low latency) rather than one assuring high throughput?09:59
tomreyntilerendering: Yes, that's correct. As Intels has a server CPU line, Xeon, AMD also has one, called Epyc.10:37
tomreynthis said, there are also companies running datacenters with intel desktop/workstation cpu based dedicated 'servers', since this can still be worthwhile for the cheaper segment.10:40
tomreynoh, you cross-posted :-/10:42
tilerenderingtomreyn: I wonder whether I should migrate my proxmox on ryzen to proxmox on xeon10:59
tomreyntilerendering: i don't think i can help you make this decision.11:01
dannfwhere can i find docs on netbooting subiquity? google wasn't much help there18:57
albechjust spun up a backup of an old server. Within that snapshot is the system disk which is accessible, but there is also another disk which is marked data on the hypervisor.. That particular disk doesnt appear in fstab and when i try to mount it i get an error that it is: wrong fs type, bad option, bad superblock. Looking and the disk in fstab doesnt reveal anything else.. Could this some encrypted partition.?21:31
albechNote that I had to chroot into the system and create a new sudo user21:32
albechand I didnt build the system initially21:32
albechwent through old history and no signs of mounting this other drive21:33
ahasenackwhat does fdisk -l say about that disk?21:50
=== RoyK^ is now known as RoyK
albechahasenack: https://paste.ubuntu.com/p/bC4xqbPTsn/23:46

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!