4400+ cron processes causes server crash ...

[Marc G. Fournier]

Evening all ...

  One of my servers just crashed with the "pmap_new_proc: u_map allocation
failed" ...

  Looking at a ps of the vmcore file, I find:

neptune# awk '{print $11}' /tmp/ps.crash | sort | uniq -c
   1 (Xvfb)
   1 (aac0aif)
   1 (adjkerntz)
   1 (analog)
   1 (bufdaemon)
4412 (cron)
   8 (csh)
  84 (ctl_cyrusdb)
   3 (ctl_deliver)
   1 (emacs)
   1 (find)
   1 (getty)
   1 (grep)
 313 (httpd)
  37 (imapd)
  23 (imapproxy)
  15 (inetd)
   1 (init)
   1 (ipaudit)
   5 (java)
 194 (lmtpd)
 479 (master)
   1 (mountd)
   1 (mysqld)
   1 (named)
   5 (nfsd)
   7 (nsd8x)
   1 (pagedaemon)
  13 (perl)
   2 (pine)
   4 (pipe)
  31 (pop3d)
   1 (portmap)
 280 (postgres)
   1 (ps)
   4 (python2.1)
  34 (qmgr)
   1 (rpc.statd)
   2 (rsync)
   1 (rwhod)
   1 (scp)
   4 (screen)
  14 (sh)
   3 (ssh)
  61 (sshd)
   1 (swapper)
   1 (syncer)
  40 (syslogd)
  18 (tcsh)
  11 (timsieved)
   1 (upclient)
   1 (vmdaemon)
   1 (vnlru)
   1 COMMAND

  Is there any way of finding out what jails "owned" those cron jobs
*after* the crash?  I know I can find out on a running systems using
proc/*/status, but what about after the server has crashed? :(

  On a 'normally running server', I see:

neptune# ps aux | grep cron | grep sbin | wc -l
      40