Re: signal 13 (SIGPIPE) no_active_lock errors

[Mark . Burgess]

Can this have something to do with processes hanging when binaries
change on disk (e.g. during upgrades) ?

M


On 25 Nov, Kurt Lieber wrote:
> Of a network of approximately 2 dozen machines all running 2.1.0p1, I am
> seeing errors similar to the following on ~3 of those machines:
> 
> -------------------------------------------------------------------------
> cfengine:adm01: Couldn't send
> cfengine:adm01: send
> cfengine:adm01: Received signal 13 (SIGPIPE) while doing [no_active_lock]
> cfengine:adm01: Logical start time Tue Nov 25 01:29:58 2003
> cfengine:adm01: This sub-task started really at Tue Nov 25 01:30:00 2003
> 
> cfengine:adm01: Couldn't send
> cfengine:adm01: send
> cfengine:adm01: Received signal 13 (SIGPIPE) while doing [no_active_lock]
> cfengine:adm01: Logical start time Tue Nov 25 01:29:58 2003
> cfengine:adm01: This sub-task started really at Tue Nov 25 01:30:00 2003
> -------------------------------------------------------------------------
> 
> This error only seems to show up on file copies and even then, it only
> appears on some of them.  (no consistent patterns)  If it fails on one run,
> it will often succeed without problems the next run, even though all the
> files in question are *identical*
> 
> I have verified that authentication is working fine and that all the
> clients are sync'd via NTP.
> 
> There are a few posts in the archives about this:
> 
> http://mail.gnu.org/archive/html/help-cfengine/2003-01/msg00048.html
> http://mail.gnu.org/archive/html/help-cfengine/2002-10/msg00035.html
> http://mail.gnu.org/archive/html/help-cfengine/2002-04/msg00087.html
> 
> But no solutions.
> 
> Any suggestions?
> 
> --kurt



~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Work: +47 22453272            Email:  Mark.Burgess@iu.hio.no
Fax : +47 22453205            WWW  :  http://www.iu.hio.no/~mark
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~