| Subject | Re: [firebird-support] Firebird Hanging - Need suggestions ??? |
|---|---|
| Author | Helen Borrie |
| Post date | 2006-01-21T21:59:31Z |
At 08:15 PM 21/01/2006 +0000, you wrote:
on RedHat 9 by an environment setting (see release notes). Did someone
perhaps apply an upgrade without checking whether it affected the
environment variables?
Databases: 14
Max users (based on max connections allowed to DB Pools): 70
Average Users (Based on Connections reported in DatabaseWorkbench): 55
=================
You have users (at server level) and requests being processed for 14 databases.
In SS, the number of users is always going to be less than the number of
process threads. I'd look at this limit of 70 connections. The number of
connection requests wouldn't need to exceed the average by much to hit your
connection limit.
Run a ps -aux |grep fbserver to count the threads.
users connected to any databases then obviously the resource is
*there*. Hence my first guess that, when you see network error 2, it's
your connection pool limit kicking in.
If it has never bitten before, look around new employees and see whether
there's someone using the Off/On switch to close sessions. Busted
connections don't just evaporate - they have to time out and clean up.
./heLen
>I've had Firebird running non-stop on this machine for months..Does the client get any message?
>
>Suddenly it is hanging, twice in the last several days.. (I define
>hanging as the inablility to complete any connection)
>Some basic background info..Very old. Fb 1.5 has had a lot of fixes in the past two years.
>===================
>
>Firebird Version installed FirebirdSS-1.5.0.4290-0.i686.rpm
> (Note: It is the ONLY application running on this machine.)
>Redhat 9 (with last updates available)A threading problem? For that build of Firebird you need to disable NPTL
> Kernel: kernel-2.4.20-30.9.i686.rpm
on RedHat 9 by an environment setting (see release notes). Did someone
perhaps apply an upgrade without checking whether it affected the
environment variables?
>HardwareDatabase Usage:
> Processor: Intel P4 2.8 GHz
> Memory: 1 Gig
> Free Disk Space on Data Partition: 99 Gigs
>
>
>
>Obviously, I'm very concerned..<sigh>
>
>Can anyone help me diagnose this problem.. What to check for..etc..
>
>Have I hit some limit of databases or users..???
Databases: 14
Max users (based on max connections allowed to DB Pools): 70
Average Users (Based on Connections reported in DatabaseWorkbench): 55
=================
You have users (at server level) and requests being processed for 14 databases.
In SS, the number of users is always going to be less than the number of
process threads. I'd look at this limit of 70 connections. The number of
connection requests wouldn't need to exceed the average by much to hit your
connection limit.
Run a ps -aux |grep fbserver to count the threads.
>microps-185 (Client) Fri Jan 20 00:03:00 2006Means a resource the client is connecting to isn't available. If you have
> INET/inet_error: receive in try_connect errno = 2
users connected to any databases then obviously the resource is
*there*. Hence my first guess that, when you see network error 2, it's
your connection pool limit kicking in.
If it has never bitten before, look around new employees and see whether
there's someone using the Off/On switch to close sessions. Busted
connections don't just evaporate - they have to time out and clean up.
./heLen