After an upgrade to QAS 4.0.1, cron starts reporting errors like the one below when many cron jobs are run at the same time. The owner of the failed cron job may be a QAS user or a local user. The error "ts=13" indicates a broken pipe.
> root 18582 c Mon Feb 14 16:00:00 2011
< root 18582 c Mon Feb 14 16:00:00 2011 ts=13
On Solaris, when an IPC hits the queue when it is full, the request could re-connect incorrectly and write on an invalid pipe, leading to SIGPIPE. This was seen when cron kicked off many requests at once.
This defect was fixed in QAS 220.127.116.11; upgrading to this version or higher will solve the problem.