RPM buglet, postgres-devel

-----BEGIN PGP SIGNED MESSAGE-----

I have finally needed to use ipcclean. I recently built HaruspeX 4.0,
an application for managing pictures which uses Postgres, and while
asking to delete a table, it hung. I had the app in the debugger, and
I found it was hung on the database call itself. So, after waiting a
couple of hours, I decided to bounce postgres by running the init
script. HaruspeX dutifully reported that the backend had closed the
connection. However, postgres did not come back up. I found two
processes running as postgres:

532 root> ps -efw | grep post
postgres 26171 1 0 00:02 ? 00:01:04 /usr/bin/postgres localhost roland family idle
postgres 26188 1 0 00:03 ? 00:00:00 /usr/bin/postgres localhost roland family idle
root 6681 20510 0 21:20 pts/2 00:00:00 grep post

When I attempted to run pgctl directly, it reported a problem with
semaphore space, so I tried to run ipcclean. It kept reporting I had
a postmaster running when I couldn't find it. Finally, I had a look
at the script. Oops, it has this line:

if ps x | grep -s 'postmaster' >/dev/null 2>&1 ; then

But consider:

557 root> ps x | grep -s 'postmaster'
7084 pts/2 S 0:00 grep -s postmaster

So, it keeps seeing its test instead of the real postmaster. How about

ps x | grep -s postmaster | grep -v grep

instead?

roland
- --
PGP Key ID: 66 BC 3B CD
Roland B. Roberts, PhD Unix Software Solutions
roberts@panix.com 76-15 113th Street, Apt 3B
rbroberts@acm.org Forest Hills, NY 11375

-----BEGIN PGP SIGNATURE-----
Version: 2.6.3a
Charset: noconv
Comment: Processed by Mailcrypt 3.5.4, an Emacs/PGP interface

iQCVAwUBOYoeLOoW38lmvDvNAQGFsAP+IHgR7bCai2W8mV9eJQvMMid29ulz4n0T
gHqVDE7hO7KWR+Lrd3KDmCQvOtEGVQazZXwp74SWBtWt1FBMUaLa8HzN3uhO4PPB
77jssDJJchC7ivNo49cxkRu+rCDK/u0GX7iw+sZQSHWwgQnzx69+9/XmhLkT3bUp
e8LkSp7jly8=
=kwCA
-----END PGP SIGNATURE-----