Oracle FAQ | Your Portal to the Oracle Knowledge Grid |
![]() |
![]() |
Home -> Community -> Usenet -> c.d.o.server -> finding how a process got started
Platform: Oracle 10.2.0.2.0 on HP-UX B.11.11
Posting to comp.unix.admin, comp.databases.oracle.server
A few weeks ago we began noticing that the .sh_history file for 'oracle' was being flooded with entries like this:
echo Start Command; LC_ALL=C;export LC_ALL;netstat -i; echo Finish Command
It was very annoying to be working and suddenly have your last command pushed way back up the line by this, but no one could account for it.
Then, yesterday while running down some active procedures (unrelated issue) I noticed this:
oracle> ps -ef |grep exp|grep -v grep
oracle 2803 2792 0 07:38:45 ? 0:00 sh -c LC_ALL=C;export
LC_ALL;vmstat 15
Hmm, looks very much like what we are seeing in the history file. Chasing the lineage of the pids, I was able to construct this chain:
root 985 1 0 Jun 21 ? 40:08 /opt/ssh/sbin/sshd root 2789 985 0 07:38:44 ? 0:00 sshd: oracle_at_notty <snip several children of 985> root 2792 985 0 07:38:44 ? 0:00 sshd: oracle_at_notty oracle 2803 2792 0 07:38:45 ? 0:00 sh -c LC_ALL=C;export LC_ALL;vmstat 15 oracle 2808 2803 0 07:38:45 ? 0:00 vmstat 15
Of interest are the first entry and the last three ... pids 1, 985, 2792, 2803, and 2808.
I took this to our SA, hoping he could tell me where 985 came from and how it came to spawn 2792, but he hadn't a clue.
Any ideas on how we can chase this down? Received on Tue Aug 14 2007 - 08:36:16 CDT
![]() |
![]() |