*** juergbi has quit IRC | 02:01 | |
*** laurenceurhegyi has quit IRC | 02:18 | |
*** gitlab-br-bot has quit IRC | 02:18 | |
*** edcragg has quit IRC | 02:18 | |
*** edcragg has joined #baserock | 02:20 | |
*** laurenceurhegyi has joined #baserock | 02:26 | |
*** juergbi has joined #baserock | 02:29 | |
*** gitlab-br-bot has joined #baserock | 02:33 | |
*** ctbruce has joined #baserock | 07:32 | |
*** noisecell has joined #baserock | 07:46 | |
*** toscalix has joined #baserock | 07:58 | |
*** ssam2 has joined #baserock | 09:08 | |
*** ChanServ sets mode: +v ssam2 | 09:08 | |
*** CTtpollard has joined #baserock | 09:17 | |
gitlab-br-bot | ybd: issue #263 ("YBD fails if the artifacts ( YBD_artifacts) are in another volume") changed state ("opened") https://gitlab.com/baserock/ybd/issues/263 | 10:18 |
---|---|---|
jjardon | hi! any idea if ^ is expected or is actually a bug in ybd? | 10:19 |
ironfoot | hm.. I think that's expected | 10:36 |
ironfoot | and I think a workaround would be to have the tmp folder in the same volume | 10:37 |
jjardon | I can not have it in the same volume for my use case | 10:43 |
jjardon | Seems the problem is that ybd use hard links when staging; any idea why is doing this? would be possible to use soft links instead? Or maybe ybd should check and move the artifact to another location if it detects they are in another volume? | 10:45 |
ssam2 | the point is to avoid copying huge artifacts around | 10:46 |
ssam2 | i guess the check could be removed, but then builds would be super slow if things weren't on the same volume | 10:47 |
jjardon | thre is not check; ybd fails when trying to create the hard links | 10:47 |
ssam2 | ah right | 10:47 |
ssam2 | i've no idea whether we could use softlinks, i've never seen anything do that | 10:48 |
ssam2 | the fact that ostree has always just used hardlinks suggests to me that it's the only way | 10:48 |
ironfoot | ybd unpacks the artifacts in the same folder as the compressed ones. If we make ybd unpack them in an special folder in tmp, (or whatever folder is used for staging) it will fix your issue | 10:52 |
*** ctbruce has quit IRC | 11:24 | |
*** ctbruce has joined #baserock | 11:51 | |
*** tiagogomes has quit IRC | 13:40 | |
paulsherwood | jjardon: the hardlinks were copied from morph iirc | 14:07 |
paulsherwood | and i have a vague recollection that there was a problem with softlinks | 14:08 |
ironfoot | morph used to put the unpacked artifacts in <tmpdir>/chunks | 14:11 |
ironfoot | (avoiding this issue) | 14:12 |
paulsherwood | did that involve an extra step? i'm struggling to recall why i would have done anything different for this | 14:14 |
paulsherwood | jjardon: why does tmp have to be on a different volume for your usecase? | 14:16 |
jjardon | tmp no, the ybd artifacts folder | 14:16 |
jjardon | they are stored in a different volume | 14:17 |
paulsherwood | why can't tmp and artifacts be on the same volume? | 14:17 |
paulsherwood | (eg move tmp to your artifacts volume?) | 14:18 |
jjardon | yeah, let me try that | 14:18 |
*** ctbruce has quit IRC | 15:23 | |
jjardon | paulsherwood: that worked, thanks! | 16:27 |
*** CTtpollard has quit IRC | 17:02 | |
*** CTtpollard has joined #baserock | 17:02 | |
*** CTtpollard has quit IRC | 17:04 | |
*** persia has quit IRC | 17:18 | |
ironfoot | Heh, I suggested the same thing :P | 17:20 |
*** ssam2 has quit IRC | 17:25 | |
*** toscalix has quit IRC | 17:52 | |
*** inara has quit IRC | 18:23 | |
*** jjardon has quit IRC | 18:23 | |
*** inara has joined #baserock | 18:25 | |
*** jjardon has joined #baserock | 18:31 | |
*** ctbruce has joined #baserock | 18:33 | |
*** rdale_ct has quit IRC | 18:34 | |
*** ctbruce has quit IRC | 19:25 | |
*** jjardon has quit IRC | 20:01 | |
*** jjardon has joined #baserock | 20:06 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!