Yes, they do go back to zero after some minutes. Regarding Dale's comments: I cannot see any problem with the SIP scenarios between the load test agents (call sources and destinations) and SipX (which does not mean that there is not any problem :-)).
The result means to me that SipX under load requires quite an amount of memory (beside CPU load). Is this conclusion correct? Regards, Gabor -----Original Message----- From: Kathleen Eccles [mailto:[email protected]] Sent: 27 November 2009 01:19 To: Gabor Paller Cc: sipx-users Subject: RE: [sipx-users] removeOldTransactions On Thu, 2009-11-26 at 17:35 +0000, Gabor Paller wrote: > I can see this (graph attached). The numbers you see in the graph are > the original transaction numbers (removeOldTransactions deleting x of y, > the graph shows the y value). > > The number of transactions does stabilize around 13000 (the decrease of > transactions at the end of the graph is when the load test was shut > down). Meanwhile, however, the memory footprint of the sipXproxy process > goes up to 63 mega as displayed by the SZ column of ps -Afl command > (initial footprint: about 10 mega). The load was about 2 call per sec. > > Can you see anything in these data that seems abnormal? Or is it how it > should work? > Does the number of transactions return to 0 after the test is stopped? If so, nothing sticks out as wrong. If not, we will need more data. -Kathy _______________________________________________ sipx-users mailing list [email protected] List Archive: http://list.sipfoundry.org/archive/sipx-users Unsubscribe: http://list.sipfoundry.org/mailman/listinfo/sipx-users sipXecs IP PBX -- http://www.sipfoundry.org/
