Ubuntu Build Service jobs stuck in the build queue
Bug #956739 reported by
Fathi Boudra
This bug affects 1 person
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Linaro CI |
Fix Released
|
Medium
|
Deepti B. Kalakeri |
Bug Description
Since yesterday (2012-03-15), UBS[1] jobs are stuck in the build queue because they can't get a node (they're all offline).
It's seems caused by the main Ubuntu archive for Precise. As a result, we can't download packages and setup the node.
Could you take a look? Also, ensure we're proactive on those connectivity issues and get notified (timeout available?).
[1] https:/
Changed in linaro-ci: | |
status: | New → Triaged |
importance: | Undecided → High |
Changed in linaro-ci: | |
importance: | High → Medium |
Changed in linaro-ci: | |
assignee: | nobody → Deepti B. Kalakeri (deeptik) |
milestone: | none → 2012.05 |
Changed in linaro-ci: | |
milestone: | 2012.05 → 2012.06 |
Changed in linaro-ci: | |
status: | Fix Committed → Fix Released |
To post a comment you must log in.
Ah! Fathi. The reason for your builds getting queued is not because of the packages non availability this time.
Its because nodes are not provisioned because the Instance Cap of 20 instances were reached
Here is the jenkins log information: plugins. ec2.EC2Cloud provision plugins. ec2.EC2Cloud provision plugins. ec2.EC2Cloud provision
INFO: Instance cap reached, not provisioning.
Mar 16, 2012 7:50:07 AM hudson.
INFO: Instance cap reached, not provisioning.
Mar 16, 2012 7:50:17 AM hudson.
INFO: Instance cap reached, not provisioning.
Mar 16, 2012 7:50:17 AM hudson.
INFO: Instance cap reached, not provisioning.
There seems to be high usages of ec2 instances at the same time because of which we are facing this problem.
I could only wish we increased the instance cap of ec2 instances now across linaro from 20 to something higher.
Until we have this cap limit we will tend to get such problems often.
Another way to reduce this is provisioning your builds at different intervals so that they dont get stuck in the queue at the same time. Between I try to use more executors on the instances so that more parallel builds can work on the same slave.
Hope that will help to solve the problem to some extent.
Could you reduce the severity of the problem to medium?
Thanks!!!
Deepti.