juergh | akselmo, creating the patch from the code snippet in the ticket is the easy part. coming up with a justification why it's needed not so much without knowing anything about it ;-) | 07:25 |
---|---|---|
=== Beret- is now known as Beret | ||
=== sashal_ is now known as sashal | ||
nemo | Hey folks. I encountered rpeated crash in Ubuntu 20.04 LTS when they reboot one of the remote CIFS mounts that looks a lot like https://www.suse.com/support/kb/doc/?id=000019780 | 13:32 |
nemo | https://github.com/openSUSE/kernel/commit/a52930353eaf443489a350a135c5525a4acbbf56#diff-365e109294d97a19e9458fa962d35d39ee87651aa464ad34c68606d81783c170 | 13:32 |
ubottu | Commit a529303 in openSUSE/kernel "cifs: handle empty list of targets in cifs_reconnect()" | 13:32 |
nemo | I was wondering if this never got pushed upstream | 13:32 |
nemo | Could someone help me locate the branch for 20.04 LTS 5.40.0-80 x86_64 so I could look over that section? | 13:33 |
nemo | fs/cifs/connect.c | 13:33 |
nemo | symptoms are super annoying at present since mounts just hang and rebooting seems only fix. | 13:34 |
nemo | (our current workaround is just to leave the windows mounts active only briefly on Ubuntu 20.04) | 13:34 |
mhcerri | hi, nemo. you can find the source here: https://git.launchpad.net/~ubuntu-kernel/ubuntu/+source/linux/+git/focal/log/?h=Ubuntu-5.4.0-80.90 | 14:01 |
nemo | mhcerri: thank you | 14:03 |
mhcerri | nemo, I think we have this commit in the version in proposed as by https://bugs.launchpad.net/bugs/1935833 | 14:03 |
ubottu | Launchpad bug 1935833 in linux (Ubuntu Focal) "linux-azure CIFS DFS oops" [Medium, Fix Committed] | 14:03 |
nemo | mhcerri: oooh even better | 14:04 |
nemo | good to have someone experieinced looking at it | 14:04 |
nemo | mhcerri: so it'll be in backports shortly? | 14:04 |
mhcerri | let me check the proposed version to confirm | 14:04 |
nemo | mhcerri: interesting. reported only recently. The opensuse one was fixed last summer | 14:04 |
nemo | I wonder how often things like this fail to make it back upstream ☹ | 14:05 |
nemo | hm. | 14:07 |
nemo | I wonder how similar the OpenSUSE code is to the ubuntu | 14:07 |
nemo | let's see ☺ | 14:07 |
mhcerri | nemo, yeah. this commit in included in 5.4.0-81.9 | 14:07 |
mhcerri | https://git.launchpad.net/~ubuntu-kernel/ubuntu/+source/linux/+git/focal/log/?h=Ubuntu-5.4.0-81.91&qt=grep&q=cifs | 14:08 |
nemo | man. this was driving us bonkers for like a month | 14:08 |
nemo | moutn hangs, reboot, mount hangs, reboot | 14:08 |
nemo | finally made a sudo script to load mounts on demand as needed | 14:09 |
nemo | but good to know. thanks! | 14:09 |
mhcerri | nemo, if you want you can try to install the version in focal-proposed to confirm it fixes it. this version is scheduled to be released next week | 14:09 |
nemo | mhcerri: could do that in test perhaps | 14:10 |
nemo | Loool | 14:10 |
nemo | Signed-off-by: Paulo Alcantara (SUSE) <pc@cjr.nz> | 14:10 |
nemo | w/ author date of last year | 14:10 |
nemo | ok. so just the opensuse one. very very late... and somehow missing the 20.04 release. just an oversight I suppose | 14:11 |
nemo | so. yay! | 14:11 |
nemo | thanks again. | 14:11 |
nemo | mhcerri: so. sorry. I've never done this before and don't want to screw up the server, even if it is just our "cmtest" server | 14:15 |
nemo | mhcerri: I just do apt -t focal-proposed update linux-kernel ? | 14:15 |
nemo | then reboot and wait to see if the mount dies again? | 14:15 |
nemo | oh. I guess I probably need to make sure the apt list includes proposed. lessee | 14:16 |
mhcerri | nemo, you can find the details in this link: https://wiki.ubuntu.com/Testing/EnableProposed | 14:16 |
nemo | thanks | 14:16 |
nemo | https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/log/fs/cifs/connect.c hmmmm checked in last august. what the heck | 14:19 |
nemo | how did this not make it into ubuntu 20.04! | 14:19 |
nemo | https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/fs/cifs/connect.c?id=a52930353eaf443489a350a135c5525a4acbbf56 | 14:19 |
ubottu | Commit a529303 in kernel/git/torvalds/linux.git "cifs: handle empty list of targets in cifs_reconnect()" | 14:19 |
nemo | oh wait. I'm an idiot. ofc it wouldn't be in 20.04 - last year, the plague year is melding in my mind with this one | 14:20 |
nemo | it just took a long time to get pulled into backports 😃 | 14:20 |
mhcerri | nemo, that was only included in 5.9 and since it doesn't have a fixes tag or cc:stable in it, it is not easy to "find" such fixes until we have a bug report and we start to investigate it | 14:21 |
nemo | right | 14:21 |
nemo | mhcerri: naw. it was me mixing up 21.04 and 20.04 ☺ | 14:21 |
mhcerri | hehe | 14:21 |
=== sarnold_ is now known as sarnold | ||
=== lan3y is now known as Laney | ||
akselmo | juergh: heh true. i'll keep talking with the peeps about this on github and hopefully we get somekind of patch for this one day. thanks for the help everyone! | 16:15 |
=== kees_ is now known as kees |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!