=== zyga is now known as zyga-afk | ||
=== retoaded_afk is now known as retoaded_gftd | ||
=== zyga-afk is now known as zyga | ||
=== _salem is now known as salem_ | ||
mvo | jibel: "+ [jibel] provide a option to run the upgrade with "eatmydata" (for the upgrade+tester): DONE" \o/ - where can I merge it? | 12:28 |
---|---|---|
mvo | jibel: and did I mention that you rock? lp:~jibel/update-manager/AutoUpgradeTester-universe ftw! | 12:32 |
jibel | mvo, re eatmydata, I didn't proposed a merge, the gain was 10s on a 13 minutes upgrade. do you think it's worth it ? | 12:40 |
jibel | mvo, we'd better preallocate disk space and use tmpfs | 12:40 |
mvo | hm, that sounds like something went wrong and it was not used? | 12:41 |
mvo | if thats all, then its indeed not worth it | 12:41 |
jibel | mvo, nope, I think kvm spend more time extending the disk than dkpg doing fsyncs. | 12:43 |
=== retoaded_gftd is now known as retoaded | ||
* mvo nods | 12:47 | |
=== zyga is now known as zyga-food | ||
=== bladernr_afk is now known as bladernr_ | ||
=== zyga-food is now known as zyga | ||
jibel | mvo, to catch deconf prompts during automated upgrades, I use the editor frontend and set EDITOR="cat>logfile" | 16:38 |
jibel | mvo, but the upgrade-tester uses DistUpgradeViewNonInteractive.py frontend which forces DEBIAN_FRONTEND to ... noninteractive | 16:38 |
jibel | mvo, do you think of any other way to do that, or should I change DistUpgradeViewNonInteractive.py to no change the frontend if it already exists in the environment ? | 16:39 |
mvo | jibel: please make DistUpgradeViewNonInteractive.py setup the EDITOR and DEBIAN_FRONTEND=editor envs | 16:40 |
mvo | jibel: that will still be non-interactive, right? but also log whats going on | 16:40 |
mvo | jibel: alternatively we could make it a config option | 16:40 |
jibel | mvo, right, it is still non-interactive, setting editor to cat is like pressing enter to all questions | 16:41 |
mvo | jibel: aesome, so it fits into there | 16:42 |
jibel | mvo, great, thanks | 16:42 |
=== bladernr_ is now known as bladernr_afk | ||
=== kalosaurusrex is now known as albriga | ||
=== albriga is now known as albrigha | ||
GrueMaster | stgraber: I can't login to the iso.qa.ubuntu.com site. favicon.ico missing. | 18:16 |
stgraber | GrueMaster: confirmed ... not quite sure what's going on there, having a look now | 18:17 |
stgraber | GrueMaster: better now? | 18:18 |
stgraber | I blamed the cache as usual, wiped it and turned it off for now, seems to work here | 18:19 |
GrueMaster | Yep, that worked for me. | 18:19 |
GrueMaster | One of these days I need to figure out a way to automate marking the daily core images as tested. I have a jenkins job that tests all 4 images daily. | 18:21 |
stgraber | GrueMaster: the API works fine for that now. Want an example script? | 19:21 |
stgraber | I'm pushing the daily upgrade tests this way | 19:21 |
GrueMaster | Sure. email me a copy. | 19:21 |
stgraber | GrueMaster: http://paste.ubuntu.com/807762/ | 19:28 |
stgraber | GrueMaster: you can set your API key here: http://iso.qa.ubuntu.com/qatracker/profile | 19:28 |
GrueMaster | And I create this key...how? | 19:30 |
GrueMaster | grrr. I'm hitting an acpi error on my serial console box. | 19:33 |
stgraber | GrueMaster: put whatever you want in there and use that value in the script instead of your actual password | 19:37 |
GrueMaster | Ah, ok. Kind of a pass-phrase thing. | 19:38 |
stgraber | yeah, that's to avoid having you put your SSO password cleartext in a script :) | 19:38 |
stgraber | to test you can either go look on the tracker or chck the value of resultid, a value of -1 means something failed, otherwise you'll get the id of the result on the tracker | 19:39 |
GrueMaster | Ok. I'll experiment with it shortly. | 19:40 |
GrueMaster | Gotta fix my serial console server first. | 19:40 |
stgraber | cool, let me know if you have any problem | 19:40 |
=== salem_ is now known as _salem |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!