admin (eg Rubber ducky) tasks should have longer timeouts

[Steve McInerney]

Is exceedingly frustrating when trying to do, eg admin people merge or admin team merge (to pick 2 oops' I just got) and have the underlying system smash with a timeout oops.

https://lp-oops.canonical.com/oops.py/?oopsid=OOPS-1208ED9
https://lp-oops.canonical.com/oops.py/?oopsid=OOPS-1208EB5

Have come across this elsewhere, previously.
Just in severely irritated mode atm as am trying to do some CHR==>LOSA requests and having each task get smashed with a timeout when I click the "Make it so" button....

Cheers!
- Grumpy Steve

[Diogo Matsubara (matsubara)]

Hi Steve, these are bug 156217 and bug 162510 respectively. We need to fix those issues separately not increase the timeout threshold for admin usage. I'm closing this bug in favor of the other reports. I'll make a comment on each one to try to get the importance on those raised.

[Steve McInerney (spm)]

Diogo, hmm, I'm not so sure it's as simple as that in all cases.

eg. Trying to fulfill this Question:
https://answers.edge.launchpad.net/launchpad/+question/74120

Keeps oopsing. eg https://lp-oops.canonical.com/oops.py/?oopsid=OOPS-1265EB40
1700+ SQL queries?

[Francis J. Lacoste (flacoste)]

I agree with Diogo, the reason we have time outs on the app servers is not to piss off rubber duckies :-),l but because long running requests are bad because they tie up an app server thread, take DB resources, etc. If there is no way for a particular task to be completed through the web in a decent time out, we should offer another way to do the task (for example, recording the request and processing it offline). So, all in all, it's a case-by-case bug that needs to be fixed.