Hello,
On Sat, 24 Mar 2012 12:12:15 +0200 Fathi Boudra fathi.boudra@linaro.org wrote:
On 23 March 2012 21:26, Zach Pfeffer via RT rt@linaro.org wrote:
Fri Mar 23 19:26:01 2012: Request 363 was acted upon. Transaction: Ticket created by zach.pfeffer@linaro.org Queue: [T] Linaro Subject: android-build jenkins only looks like it can use one builder Owner: Nobody Requestors: zach.pfeffer@linaro.org Status: new Ticket <URL: https://rt.linaro.org//Ticket/Display.html?id=363 >
Would someone take a look? Mathieu is saying Jenkins can only use one builder.
Situation is back to normal ~2h after the ticket was submitted.
Well, actually I did some actions after 10 mins after seeing that info on IRC.
Closing. For reference, IS isn't in charge of this service.
Please, report a bug on Linaro Android Infrastructure for investigation: https://bugs.launchpad.net/linaro-android-infrastructure/+filebug
+1. Information Services (IS) should be contacted directly only on selected limited number of questions, like recent case of updating a file http://snapshots.linaro.org/android/binaries/ in, where Infra team doesn't access. The rest is still handled by Infra team by that link.
Such discrepancy in procedures may seem counterintuitive, and I used to see it such, and did some noise about that yet on Connect Q4.11, Danilo can prove ;-). In the end it was decided that responsibility separation is good, and we won't get around it anyway, as Linaro consists of many different teams.
Note: we have an instance cap limit shared between android-build.l.o and ci.l.o, set to 20. If the builds have been queued on android-build.l.o and 19 instances are already taken by ci.l.o, only one instance is remaining for a-b.
That's true. And recently we had: 1) more permanent instances added to runs services (9 total); 2) cases of zombie instances due to ci.linaro.org issues (wild suspicion is that it still happens, I'll be watching it closely). So, we actually had/have some instance scarcity these days. I however did audit of all running instances yesterday, and caught few more ci.linaro.org-forgotten instances. So, we should be better now. There's also discussion of increasing our global EC2 instance limit (that should be possible, but I imagine may require some paperwork with Amazon).
Cheers,
Fathi