Re: Memory leaks in orbit and/or CORBA::ORBit 0.4.3?

[Sam Couter <sam topic com au>]

Wolfgang Hottgenroth <who_l@hottis.de> wrote:
> I see in my top output on a Solaris box:
> After start:
> PID USERNAME THR PRI NICE  SIZE   RES STATE   TIME    CPU COMMAND
> 860 whottgen   1  48    0 4104K 3616K sleep   0:00  0.61% server
> 
> And roundabout 15 minutes later:
> PID USERNAME THR PRI NICE  SIZE   RES STATE   TIME    CPU COMMAND
> 860 whottgen   1  58    0 4128K 3656K sleep   0:05  0.36% server

That's only 24K difference! That's not a memory leak, that's just
resources that hadn't been allocated before you started making requests.
I'd be willing to bet that those resources are things such as malloc()
heap areas that are being reused all the time.

If you'd been watching the size, you would have seen an initial jump in
size as buffers and so on are allocated, but no more growth after that.
If it continues to grow, only then do you have a memory leak.

Just for information, I can't reproduce any growth either, and I've
hammered on my server a few million times from a bunch of different
machines.
-- 
Sam Couter          |   Internet Engineer   |   http://www.topic.com.au/
sam@topic.com.au    |   tSA Consulting      |
OpenPGP key ID:       DE89C75C,  available on key servers
OpenPGP fingerprint:  A46B 9BB5 3148 7BEA 1F05  5BD5 8530 03AE DE89 C75C

PGP signature