You may have seen a number of test failures on a few jobs, I know
specifically the Javascript test failure on Katello PRs is failing
repeatedly. This appears to be due to the Jenkins slave that some jobs
target throwing "no space left on disk errors".
Thanks, I've freed up a little space and will dig a bit further into
where the space is going.
I think it's probably just a bit limited by a 20GB root disk and having
all of the pbuilder caches.
···
On 01/09/15 03:37, Eric D Helms wrote:
> You may have seen a number of test failures on a few jobs, I know
> specifically the Javascript test failure on Katello PRs is failing
> repeatedly. This appears to be due to the Jenkins slave that some jobs
> target throwing "no space left on disk errors".
>
> If anyone has information on, or can look
> intodebian01.rackspace.theforeman.org
> to
> figure out the overall issue please let us know.
Yeah, it just can't fit. This slave is a smaller (General 1GB) flavour
compared to the regular unit test slaves (General 4GB) as it's only used
for package builds and Jenkins flyweight jobs, so only has a quarter of
the disk space.
I added a second block device to it for /var/lib/workspace which was
cheaper and easier than upgrading the flavour/size.
···
On 01/09/15 08:04, Dominic Cleal wrote:
> On 01/09/15 03:37, Eric D Helms wrote:
>> You may have seen a number of test failures on a few jobs, I know
>> specifically the Javascript test failure on Katello PRs is failing
>> repeatedly. This appears to be due to the Jenkins slave that some jobs
>> target throwing "no space left on disk errors".
>>
>> If anyone has information on, or can look
>> intodebian01.rackspace.theforeman.org
>> to
>> figure out the overall issue please let us know.
>
> Thanks, I've freed up a little space and will dig a bit further into
> where the space is going.
>
> I think it's probably just a bit limited by a 20GB root disk and having
> all of the pbuilder caches.