pcp
[Top] [All Lists]

Re: [pcp] qa/023 hanging?

To: Mark Goodwin <mgoodwin@xxxxxxxxxx>
Subject: Re: [pcp] qa/023 hanging?
From: Nathan Scott <nathans@xxxxxxxxxx>
Date: Mon, 25 Apr 2016 22:14:55 -0400 (EDT)
Cc: pcp <pcp@xxxxxxxxxxx>
Delivered-to: pcp@xxxxxxxxxxx
In-reply-to: <571EC3B0.6050600@xxxxxxxxxx>
References: <571DA64A.1080802@xxxxxxxxxx> <280513017.43174641.1461633146452.JavaMail.zimbra@xxxxxxxxxx> <571EC3B0.6050600@xxxxxxxxxx>
Reply-to: Nathan Scott <nathans@xxxxxxxxxx>
Thread-index: z48Hl62rrwZSStGOdkdjxodyiiAlLw==
Thread-topic: qa/023 hanging?

----- Original Message -----
> On 04/26/2016 11:12 AM, Nathan Scott wrote:
> ..
> >
> > FWIW, its not hanging for me either, after a couple of runs.  I am seeing
> 
> not hanging for me any more either! must have been a local system issue
> or something.

Wierd.  I noticed the only use of fuser in the test is via qa/common.check as
it uses it to seek out an unused port (_get_port) ... hmm, still not seeing a
way that could hang though, probably needed to set strace on it.  Oh well, if
we see it again we might be able to do that.

> > freshly introduced failures in 246 (derived metrics) and 4751 (locking in
> > pmNewContext) however ... see attached.
> >
> 
> 246 is failing because it's relying on malloc re-using free'd blocks in a
> predictable way. That's not very deterministic in general - I'll see if
> I can improve the filter; I think this is a test failure, not a problem with
> derived metrics.

+1 - yep, looks that way to me too, thanks Mark.

cheers.

--
Nathan

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