=== gfidente is now known as gfidente|afk | ||
=== shardy is now known as shardy_afk | ||
vans163 | Anyone know the correct way to use cloud-init with qemu? | 18:04 |
---|---|---|
vans163 | If I pass the .iso I created to qemu as -drive then remove it. I fail to boot up the second time with errors like this blk_update_request: I/O error, dev fd0, sector 0 | 18:04 |
vans163 | reading google seems to indicate maybe the cloud image expects that drive to be there | 18:04 |
powersj | vans163: I have been using something similar to: https://ubuntu-smoser.blogspot.cl/2013/02/using-ubuntu-cloud-images-without-cloud.html | 18:53 |
vans163 | powersj: doing -hda/hdb like that does not allow to pass virtio | 18:55 |
vans163 | im gonna try using -cdrom i guess | 18:55 |
vans163 | or i guess will have to keep the cloud-init.iso always mounted/attached | 18:55 |
vans163 | everytime that vm boots | 18:55 |
powersj | Well so here is what I end up using, which you can see the virtio cmd: https://gist.github.com/powersj/987069c407c5115acfd72a09ce1b9be9 | 18:57 |
vans163 | powersj: wil investigate that fully thanks, but that last command to run qemu looks identical to mine. My trouble starts is on the second boot I leave out -drive file=$SEED_FILE,format=raw,if=virtio | 18:58 |
powersj | very specific for my use-case of going through cloud configs quickly, but you can play with it | 18:58 |
powersj | ah | 18:58 |
vans163 | do you leave out this line "-drive file=$SEED_FILE,format=raw,if=virtio" on second boot and further? | 18:58 |
powersj | ah second boot? let me try | 18:59 |
vans163 | yea like make the seed.iso, attach it, boot, works. Shutdown QEMu fully, start qemu again the same main disk, BUT leave out "-drive file=$SEED_FILE,format=raw,if=virtio". It does not boot, gives those errors "blk_update_request: I/O error, dev fd0, sector 0" Do you see these at all? | 19:00 |
vans163 | By default I get 3 of those using the ubuntu xenial 16.04 cloud image | 19:00 |
vans163 | but it does not affect anything | 19:00 |
vans163 | without the $SEED_FILE theres like 10-12 of them, and Ubuntu does not boot | 19:01 |
powersj | vans163: agreed, seeing same thing | 19:02 |
vans163 | powersj: fails to boot fully too? Gets stuck after those blk_update_request errors? | 19:04 |
powersj | I wonder if it is stuck waiting for userdata or something like that | 19:04 |
vans163 | ah | 19:04 |
powersj | vans163: yes just sits there | 19:04 |
powersj | when I added the seed back in it boots as expected like you mentioned | 19:05 |
vans163 | i was going crazy at first, i thought my filesystem was corrupting the image and im using ZFS x.x | 19:05 |
powersj | ha! glad I could at least confirm it for you | 19:06 |
vans163 | yea thanks. so would this be a bug? | 19:06 |
vans163 | im not sure if its local to xenial ubuntu or if itl happen on other images | 19:06 |
powersj | I'm not sure, I don't know enough if there is some setting you would need to disable or mark on the cloud-img I'm using to allow 2nd boot | 19:06 |
vans163 | i see. I think il just run with always having the seed attached for now | 19:07 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!