Re: Load testing

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hello!

Again, the problem wasn't with the gnugk itself :) The problem was with
call generator hosts. One of them is not good enough [P4 2.4GHz, 512M
RAM]. I was able to go beyond 78 concurrent calls.
What I did is that I made 4 listening clients and 4 generating clients
(each making 30 calls). Between 30 new calls I put 5 seconds of waiting.
In this way I was achieve 112 concurrent calls in full proxy mode :)
>From load on GK I guess, it can go beyond this number, just need to find
 another/more hosts for testing. I have to add also, that GK had debug
level 5, so in production with level 0 or 1 it can be even better :)

I will keep the list informed [and maybe I will write a short report].

I have one further question:
do I have to change limits.h and fs.h in linux to get out more than 1024
 sockets or ulimit -n  is enough? I googled a bit and found that NR_OPEN
and NR_FILE is recommended to increase. While I'm not a linux low-level
guru, I don't know if it is really needed or not. Any hint?

I'm eager to know how much can be handled with one Xeon and a dual Xeon
machine....

Kind regards,
	Tamas

ps: Thank you for suggestions.

Stewart Nelson wrote:
> Hi Tamas,
> 
> Sorry, this is beyond my expertise; I can't see anything wrong.
> Perhaps Michal will recognize the problem.
> 
> Since we know that many users are handling more than 78 simultaneous
> calls, it may be worthwhile to try e.g. version 2.0.9, to see if it
> fails in the same way.
> 
> A higher trace level might show why gnugk needed to tear the call
> down.  If not, you should be able to tell from successful calls
> what should have happened next after Call Proceeding, e.g.
> opening an H.245 connection.  Then, you could look for why that
> is not happening, adding your own trace statements if needed.
> 
> Also, it's conceivable that this problem is related to the high rate
> of call establishment, and that you wouldn't see such failures at
> this traffic level in real life.  Try slowing things down, if your
> call generator allows it.
> 
> --Stewart
> 
> 
> ----- Original Message ----- 
> From: "Tamas J" <thomasj@xxxxxxxxx>
> To: <openh323gk-users@xxxxxxxxxxxxxxxxxxxxx>
> Sent: Saturday, April 09, 2005 12:58 PM
> Subject: Re:  Load testing
> 
> 
> 
>>Hello Stewart,
>>
>>thanks for the prompt response.
>>ulimit -a
>>core file size        (blocks, -c) 0
>>data seg size         (kbytes, -d) unlimited
>>file size             (blocks, -f) unlimited
>>max locked memory     (kbytes, -l) 32
>>max memory size       (kbytes, -m) unlimited
>>open files                    (-n) 8192
>>pipe size          (512 bytes, -p) 8
>>stack size            (kbytes, -s) 8192
>>cpu time             (seconds, -t) unlimited
>>max user processes            (-u) 14336
>>virtual memory        (kbytes, -v) unlimited
>>
>>large_fdset=4096
>>
>>gnugk-2.2CVS, pwlib-1.7.5, openh323-1.14.4 (I'm not able to compile
>>newer libs with gcc-3.3.4 - already wrote to the openh323 mailing list),
>>linux-2.6.11, Intel Xeon 2.8 (HT enabled), 1GB RAM, Slackware
>>
>>[RoutedMode]
>>GKRouted=1
>>H245Routed=1
>>CallSignalPort=1720
>>CallSignalHandlerNumber=24
>>RtpHandlerNumber=24
>>RemoveH245AddressOnTunneling=1
>>RemoveCallOnDRQ=1
>>DropCallsByReleaseComplete=1
>>SendReleaseCompleteOnDRQ=1
>>AcceptNeighborsCalls=1
>>AcceptUnregisteredCalls=1
>>SupportNATedEndpoints=1
>>H245PortRange=40000-50000
>>Q931PortRange=50001-60000
>>ForwardOnFacility=1
>>ConnectTimeout=180000
>>
>>[Proxy]
>>Enable=1
>>ProxyForNAT=0
>>T120PortRange=39998-39999
>>RTPPortRange=40000-60000
>>
>>There are no FW rukes on the GK box (just testing).
>>
>>What can be wrong?
>>
>>Tamas
>>
>>ps: In logs I see that GK asked for call tear down:
>>2005/04/09 20:28:32.597 4       ProxyChannel.cxx(685)   Q931    Send to
>>213.253.103.4:46472 {
>>  q931pdu = {
>>    protocolDiscriminator = 8
>>    callReference = 42037
>>    from = destination
>>    messageType = ReleaseComplete
>>    IE: Cause - Temporary failure = {
>>      80 a9                                              ..
>>    }
>>    IE: User-User = {
>>      25 80 06 00 08 91 4a 00  02 01 11 00 f8 14 b6 f6   %.....J.........
>>      9a a7 d9 11 94 7a 00 0f  1f f8 9f bd 02 80 01 00   .....z..........
>>    }
>>  }
>>  h225pdu = {
>>    h323_uu_pdu = {
>>      h323_message_body = releaseComplete {
>>        protocolIdentifier = 0.0.8.2250.0.2
>>        callIdentifier = {
>>          guid =  16 octets {
>>            f8 14 b6 f6 9a a7 d9 11  94 7a 00 0f 1f f8 9f bd
>>.........z......
>>          }
>>        }
>>      }
>>      h245Tunneling = FALSE
>>    }
>>  }
>>}
>>Before this message the last message was Call Proceeding (7.5 seconds
>>earlier).
> 
> 



-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
_______________________________________________________

Posting: mailto:Openh323gk-users@xxxxxxxxxxxxxxxxxxxxx
Archive: http://sourceforge.net/mailarchive/forum.php?forum_id=8549
Unsubscribe: http://lists.sourceforge.net/lists/listinfo/openh323gk-users
Homepage: http://www.gnugk.org/

[Index of Archives]     [SIP]     [Open H.323]     [Gnu Gatekeeper]     [Asterisk PBX]     [ISDN Cause Codes]     [Yosemite News]

  Powered by Linux