[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [gomd-devel] <STRESS TEST> Success!
From: |
rbaardman |
Subject: |
Re: [gomd-devel] <STRESS TEST> Success! |
Date: |
Tue, 11 Nov 2003 16:56:04 +0100 |
User-agent: |
Internet Messaging Program (IMP) 3.1 |
> Hi all.
>
> Just performed two stress tests (using expect-base getInfo.sh shell
> script).
>
> [Test #1 (changed)]:
> MACHINE: ATHLON XP 1700+
> REQUESTS: 2000 x "get load"
> TIME BETWEEN REQUESTS: 2 secs (sleep time)
> RESULT: PASSED
>
> [Test #2]:
> MACHINE: PIII 500
> REQUESTS: 1000 x "get load"
> TIME BETWEEN REQUESTS: 0 secs (no delay!)
> RESULT: PASSED
>
> I've noticed a strange behaviour launching the gomd in GDB.
> As GDB traps SIGNALS, closed connHandlers (==threads killed) appear in
> TOP/PS like ZOMBIE entries ("... Z ... gomd <defunct>").
Actually, I have set a timeout for every request of something like 10
milliseconds before reading, because else I was garantueed to get no
answer. This was in java2gomd
> Obviously this situation doens't appear if you launch gomd without GDB
> (and
> the massive test #2 reported _no_ defunct gomd procs/threads).
> Maybe commoncpp's Tread::exit() call is trapped by GDB so the threads
> hang.
> As they are not properly killed.
>
> I've cleaned up connHandler class (added thread Cancelation call and a
> check
> to verify if the object is correctly killed) but this problem still
> happens
> with GDB.
>
> Any idea?
>
> Byez.
>
> <rejected>
>
> PS: maybe this happen as I'm using new GDB 6.0 (notice the 0). ;)
>
>
>
> _______________________________________________
> gomd-devel mailing list
> address@hidden
> http://mail.nongnu.org/mailman/listinfo/gomd-devel
>
--
_____________________________________________________________________
Snel en voordelig ADSL nu voor iedereen bereikbaar.
Zon Breedband Budget voor EUR 14,95 per maand.
Nu tijdelijk geen aansluitkosten. Bestel snel op zonnet.nl/breedband