llvm dev - Jan 2014 - [LLVMdev] Recent buildbot failures on arxan_raphael and osu8

Hi,



I received a small flurry of buildbot emails yesterday. Most of them seem to be
repeats of previous failures and buildbot only mailed them to me because the
previous build raised an exception instead of failing. However I noticed a
couple configuration/system issues amongst them:

*         arxan_raphael is consistently hitting the 20 minute no-output timeout
when running the link command in the compile step. It occasionally passes (e.g.
build 11090) which suggests that it's only slightly over.

*         osu8 is out of disk space.

Daniel Sanders
Leading Software Design Engineer, MIPS Processor IP
Imagination Technologies Limited
www.imgtec.com<http://www.imgtec.com/>

-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20140123/8f1178a5/attachment.html>

Daniel Sanders wrote:
> Hi,
>
> I received a small flurry of buildbot emails yesterday. Most of them
> seem to be repeats of previous failures and buildbot only mailed them to
> me because the previous build raised an exception instead of failing.
> However I noticed a couple configuration/system issues amongst them:
>
> ·arxan_raphael is consistently hitting the 20 minute no-output timeout
> when running the link command in the compile step. It occasionally
> passes (e.g. build 11090) which suggests that it's only slightly over.
>
> ·osu8 is out of disk space.
Yup, I'm aware. osu8 runs everything in ram (all source and object files 
are in RAM), and the HD was never intended to take the load (small, 
slow, not redundant, and not expected to live a long life if it needs to 
spin that much). The Clang debug build -vg builder takes half the ram on 
its own at 17G, then it's running three other builders beyond that.

I had hoped that doing a make clean would fix its disproportionate 
footprint (the next largest builder is 5G), but apparently not so. I'm 
going to see whether we can buy another 32G RAM for it.

Sorry!

Nick

I suspect I'm hitting a RAM limit on that machine... either we need to
increase the timeout or I need to move this task to a machine with more
resources.

For the interim, let's bump the timeout to 2400.

Is that something you can do Daniel?

Cheers,

Joe
______________________________
Joe Abbey
Senior Director of Product Development
Arxan Technologies
jabbey at arxan.com www.arxan.com
Protecting the App Economy™

On Jan 23, 2014, at 2:37 PM, Nick Lewycky <nicholas at mxc.ca> wrote:
> Daniel Sanders wrote:
>> Hi,
>> 
>> I received a small flurry of buildbot emails yesterday. Most of them
>> seem to be repeats of previous failures and buildbot only mailed them
to
>> me because the previous build raised an exception instead of failing.
>> However I noticed a couple configuration/system issues amongst them:
>> 
>> ·arxan_raphael is consistently hitting the 20 minute no-output timeout
>> when running the link command in the compile step. It occasionally
>> passes (e.g. build 11090) which suggests that it's only slightly
over.
>> 
>> ·osu8 is out of disk space.
> 
> Yup, I'm aware. osu8 runs everything in ram (all source and object
files
> are in RAM), and the HD was never intended to take the load (small, 
> slow, not redundant, and not expected to live a long life if it needs to 
> spin that much). The Clang debug build -vg builder takes half the ram on 
> its own at 17G, then it's running three other builders beyond that.
> 
> I had hoped that doing a make clean would fix its disproportionate 
> footprint (the next largest builder is 5G), but apparently not so. I'm 
> going to see whether we can buy another 32G RAM for it.
> 
> Sorry!
> 
> Nick
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20140123/8f081ab0/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 842 bytes
Desc: Message signed with OpenPGP using GPGMail
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20140123/8f081ab0/attachment.sig>