[check_postgres] check_postgres hot_standby_delay fails

[Guillaume Lelarge]

On Fri, 2012-08-17 at 14:40 +0200, Alexander Swen wrote:
> hi (Greg/Hendrik?), 
> 
> 
> We are happy users of the check_postgres script you wrote. However, since short (we screwed one database) the hot standby check fails: 
> /usr/lib64/nagios/plugins/check_postgres --action hot_standby_delay -w 512 -c 1024 --host=server1,server2 -u nagios_user -db databasename 
> returns: 
> 
> Use of uninitialized value $slave in numeric eq (==) at /usr/lib64/nagios/plugins/check_postgres line 4710. 
> POSTGRES_HOT_STANDBY_DELAY UNKNOWN: DB "databasename" (host:server1) Invalid query returned: receive <PIPE> \n replay <PIPE> \n | time=0.03s 
> 
> 
> we tried some different syntax, same result 
> 
> /usr/lib64/nagios/plugins/check_postgres --action hot_standby_delay -w 512 -c 1024 --host=server1 --host2=server2 --dbuser=nagios_user --dbname=databasename --dbname2=databasename 
> 
> 
> Now, this happens after one of our (test-)db's was dropped and recreated on the master. And at the moment we DO have a situation where the servers are not in sync anymore. however, I would expect an error indicating that the hot standby is outside the warn/crit boundaries, not a test failure. 
> 
> 
> I have to admit that I am neither a perl guru nor a psql pro. Our psql pro is on some conference and therefor I am helping one of my co-workers. 
> 
> 
> Could you please help us out? 
> Or ís this a bug as written in an email between you guys: https://mail.endcrypt.com/pipermail/check_postgres/2012-February/000892.html 
> 

Which release of check_postgres.pl are you using? because it works for
me.


-- 
Guillaume
http://blog.guillaume.lelarge.info
http://www.dalibo.com