how to avoid a restart loop?

how to avoid a restart loop?

[Joshua N Pritikin]

[-- Attachment #1: Type: text/plain, Size: 745 bytes --]

Today I noticed that the log was filling up with:

[2005-12-03 06:29:47 UTC] LOG:  could not load root certificate file
+"/var/lib/postgresql/8.0/main/root.crt": No SSL error reported
[2005-12-03 06:29:47 UTC] DETAIL:  Will not verify client certificates.
[2005-12-03 06:29:47 UTC] FATAL:  lock file
+"/var/lib/postgresql/8.0/main/postmaster.pid" already exists
[2005-12-03 06:29:47 UTC] HINT:  Is another postmaster (PID 288) running
in data
+directory "/var/lib/postgresql/8.0/main"?

I got 1 per second!  A few megabytes later, I killed everything and
restart.  Looks OK now, but how to avoid this in the future?

Do I need a ./finish script which does:

  kill `cat /var/lib/postgresql/8.0/main/postmaster.pid`

??

Suggestions please.  ;-)

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: how to avoid a restart loop?

[Joshua N Pritikin]

[-- Attachment #1: Type: text/plain, Size: 493 bytes --]

On Sat, Dec 03, 2005 at 01:09:03PM +0530, Joshua N Pritikin wrote:
> Today I noticed that the log was filling up with:

OK, I think I know what happened.  I sent a TERM signal to the
postmaster.  On a TERM signal, the postmaster waits for all client
connections to shutdown before exiting.  Because I was confused, I think
I killed the runsv process watching the postmaster.  runsvdir probably
started a new runsv which repeatedly tried to start a new postmaster.

So I guess it was my fault.

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 189 bytes --]