| Subject | Re: [firebird-support] Re: Firebird.conf FileSystemCacheSize, Where am I going wrong? |
|---|---|
| Author | André Knappstein, Controlling |
| Post date | 2011-05-30T13:02:44Z |
AF> The case I've seem was:
AF> * Windows Server 2008 R2 (64bit, as far as I remember)* Firebird 1.5.6 Classic (32bit, of course)
AF> After some time server shows almost 100% of memory usage (8Gb).
It would be interesting to know, how many connections (child processes) are running at that time. You can use the built-in tools, but I prefer ProcessExplorer by SysInternals.
If you find more (a lot more!) processes active than you would expect, with FB 1.5.x it could be the usual case of Zombies remaining.
I forgot the details of whether it was an OS thing, or that the ODBC driver did not do something right or if it was something in FB itself, or whether it was possible to misconfigure the installation so that Zombies would become a problem.
But if I remember correctly it was something fixed for 2.0.x and not backported.
I still have a 1.5.x Classic server and it runs fine, for 50 concurrent users, but the number of Zombies keeps growing and every 4-6 weeks I just need disconnect everybody from the database and restart the parent process, then all will be fine again.
---
AF> * Windows Server 2008 R2 (64bit, as far as I remember)* Firebird 1.5.6 Classic (32bit, of course)
AF> After some time server shows almost 100% of memory usage (8Gb).
It would be interesting to know, how many connections (child processes) are running at that time. You can use the built-in tools, but I prefer ProcessExplorer by SysInternals.
If you find more (a lot more!) processes active than you would expect, with FB 1.5.x it could be the usual case of Zombies remaining.
I forgot the details of whether it was an OS thing, or that the ODBC driver did not do something right or if it was something in FB itself, or whether it was possible to misconfigure the installation so that Zombies would become a problem.
But if I remember correctly it was something fixed for 2.0.x and not backported.
I still have a 1.5.x Classic server and it runs fine, for 50 concurrent users, but the number of Zombies keeps growing and every 4-6 weeks I just need disconnect everybody from the database and restart the parent process, then all will be fine again.
---