pcp
[Top] [All Lists]

RE: qa/660 - pmwebd in an infinite loop?

To: "'Frank Ch. Eigler'" <fche@xxxxxxxxxx>
Subject: RE: qa/660 - pmwebd in an infinite loop?
From: "Ken McDonell" <kenj@xxxxxxxxxxxxxxxx>
Date: Mon, 29 Sep 2014 07:01:52 +1000
Cc: <pcp@xxxxxxxxxxx>
Delivered-to: pcp@xxxxxxxxxxx
In-reply-to: <y0mr3z2r3fl.fsf@xxxxxxxx>
References: <5420FB9C.5000004@xxxxxxxxxxxxxxxx> <y0mr3z2r3fl.fsf@xxxxxxxx>
Thread-index: AQJoqwjG6HSjMvhEGR+Z4Bmlt4z8wQJi655XmtJLwzA=
Frank,

Some more info on the looping pmwebd problem.  This is 100% reproducible on
at least one host.

> -----Original Message-----
> From: Frank Ch. Eigler [mailto:fche@xxxxxxxxxx]
> Sent: Tuesday, 23 September 2014 9:18 PM
> ...
> Sorry, not familiar at all.  It may not be relevant, but was this pmwebd
from the
> main dev branch, or the fche/pmwebd one? ...

main dev branch

> Was the
> 660 file (the test case script itself) being modified? ...

No

> ... Maybe was there a wacky
> 660.out2 -> 660 symlink in your tree, leading the
> 660 file to be appended to itself or something like that?

No nothing like that ... 660 is unchanged by the test.

This sends pmwebd into an infinite loop

curl -s -S http://localhost:44323/660

This avoids the infinite loop problem

curl -s -S http://localhost:44323/`pwd`/660

but the output is "PMRESAPI error" not the contents of the file 660

So it looks like some sort of a path issue or read() failure issue inside
pmwebd.

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