pcp
[Top] [All Lists]

Re: [pcp] FW: pmatop failing in qa

To: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Subject: Re: [pcp] FW: pmatop failing in qa
From: Nathan Scott <nathans@xxxxxxxxxx>
Date: Mon, 29 Sep 2014 00:10:46 -0400 (EDT)
Cc: PCP <pcp@xxxxxxxxxxx>
Delivered-to: pcp@xxxxxxxxxxx
In-reply-to: <003701cfd9ce$008d8a00$01a89e00$@internode.on.net>
References: <54176C9F.7020404@xxxxxxxxxxxxxxxx> <003701cfd9ce$008d8a00$01a89e00$@internode.on.net>
Reply-to: Nathan Scott <nathans@xxxxxxxxxx>
Thread-index: AQEKyLvnJWR1kMAuHeDmXuJWQyehcJ2eCi7QyAU8EKQ=
Thread-topic: pmatop failing in qa
Hi Ken,

----- Original Message -----
> This one is still unresolved.
> 
> git blame suggests Stan or Nathan may be able to help.
> 
> [...]
> Attached are two 722.full files ... the big one is from a pass.  The short
> one (722.full.fail) is from a failed run.
> 
> Could someone who knows and cares, please take a look?
> 

It looks like we're simply bailing out part way through the pmatop invocation
and hence all expected output beyond a certain point of each invocation is
missing.  I'd expect a python exception is being generated there but it looks
like pmatop squashes all exceptions other than some special KeyboardInterrupt
exception handling (which wont be happening here of course).

Looking at your earlier overall QA status mail, it looks like this failure is
happening on three hosts:

Host      bo bl bv gr 00 01 02 03 04 05 07 10 11 12 14 18 19 21 22 23 24
Test %bad                                                                Test 
QA groups
 722  14%                             X  X     X                          722 
python

2014-09-04  639  607   32   98|vm05 vm05 PCP 3.10.0 i486 Gentoo 2.0.3 (not 
listed on gentoo.org)
2014-09-03  667  651   16   70|vm07 PCP 3.9.10 x86_64 Debian 6.0.9 (squeeze - 
oldstable)
2014-09-03  667  648   19   70|vm11 PCP 3.9.10 i686 Debian 6.0.9 (squeeze - 
oldstable)


Not 100% sure about the Gentoo one (tho it looks old too?), but those Debian
versions will have very old python versions.  I have it on my to-do list for
this week to revisit the python changes from while I was away, which at this
stage I'm punting will result in dropping support of pre-2.6 python versions.
That will in turn resolve this failure without us having to dig further into
whatever these old python versions are complaining about in pmatop code.

cheers.

--
Nathan

<Prev in Thread] Current Thread [Next in Thread>