From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
---|---|
To: | Massimo Dal Zotto <dz(at)cs(dot)unitn(dot)it> |
Cc: | hackers(at)postgreSQL(dot)org (PostgreSQL Hackers) |
Subject: | Re: [HACKERS] How do I get the backend server into gdb? |
Date: | 1999-04-29 22:33:46 |
Message-ID: | 18123.925425226@sss.pgh.pa.us |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
Massimo Dal Zotto <dz(at)cs(dot)unitn(dot)it> writes:
> The -W option is passed to the backend which sleeps 15 seconds before doing
> any work. In the meantime you have the time to do a ps, find the backend pid
> and attach gdb to the process.
> Obviously you can't do that in a production environment because it adda a
> fixed delay for each connection which will make your users very angry.
Since it's a -o option, I see no need to force it to be used on every
connection. Instead start psql with environment variable
PGOPTIONS="-W 15"
or whatever you need for the particular session. The PGOPTIONS are sent
in the connection request and then catenated to whatever the postmaster
might have in its -o switch.
(BTW, it might be a good idea to go through the backend command-line
switches carefully and see if any of them could be security holes.
I'm feeling paranoid because of Matthias Schmitt's unresolved report...)
regards, tom lane
From | Date | Subject | |
---|---|---|---|
Next Message | Tom Lane | 1999-04-29 22:39:28 | Re: [HACKERS] PLpgSQL Stat Problem |
Previous Message | Michael J Davis | 1999-04-29 20:46:11 | Problem with insert into select from using aggregation |