Hi Arun, To decouple the registration rate from the multiplier, you would alter this area as you see fit: ----- # Calculate calls/registers per second based on the input RPS = RE_REG_PER_HOUR * args.subscriber_count * args.multiplier / 3600 CPS = OUTGOING_BHCA * args.subscriber_count * args.multiplier / 3600 -----
Looking at Rob's mail, it should read calls/hour, not calls/second, as mentioned here: https://clearwater.readthedocs.io/en/latest/Clearwater_stress_testing.html#running-stress-ims-core-only Note that the multiplier is only going to get you so far - as Rob pointed out, it will probably fail over around one reg/call per second per subscriber. You'll need to add more subscribers to generate more load, and potentially nodes to handle it. Good luck! Cheers, Will From: Clearwater [mailto:[email protected]] On Behalf Of Arun Lal Sent: 27 April 2018 06:24 To: [email protected] Subject: Re: [Project Clearwater] Regarding Increasing #calls using stress testing Hi William, I have a look the Rob's answer and verify that in my setup, according to Rob: ->When using the run_stress script, each emulated subscriber generates 1.3 calls/second (50% incoming, 50% outgoing). This can be increased with the ‘multiplier’ option (e.g. --multiplier 10 means 13 calls/second) But here I use --multiplier=1000, so according to rob it would be 1300 call/sec. But it is around 16 call/sec (below in sipp-output format). So my question was, How can I increase the Call/sec may be 10000 call/sec or more (is it possible using stress testing clearwater?) []ubuntu@stress:~$ /usr/share/clearwater/bin/run_stress rags.mf 100 1 --multiplier=1000 Starting initial registration, will take 1 seconds Initial registration succeeded Starting test Test complete Elapsed time: 00:01:07 Start: 2018-04-27 18:10:06.778960 End: 2018-04-27 18:11:14.140203 Total calls: 1083 Successful calls: 1082 (99.9076638966%) Failed calls: 1 (0.0923361034164%) Unfinished calls: 0 Retransmissions: 0 Average time from INVITE to 180 Ringing: 124.0ms # of calls with 0-2ms from INVITE to 180 Ringing: 0 (0.0%) # of calls with 2-10ms from INVITE to 180 Ringing: 0 (0.0%) # of calls with 10-20ms from INVITE to 180 Ringing: 0 (0.0%) # of calls with 20-50ms from INVITE to 180 Ringing: 5 (0.461680517082%) # of calls with 50-100ms from INVITE to 180 Ringing: 409 (37.7654662973%) # of calls with 100-200ms from INVITE to 180 Ringing: 582 (53.7396121884%) # of calls with 200-500ms from INVITE to 180 Ringing: 81 (7.47922437673%) # of calls with 500-1000ms from INVITE to 180 Ringing: 6 (0.554016620499%) # of calls with 1000-2000ms from INVITE to 180 Ringing: 0 (0.0%) # of calls with 2000+ms from INVITE to 180 Ringing: 0 (0.0%) Failed: call success rate 99.9076638966% is lower than target 100.0%! Total re-REGISTERs: 3333 Successful re-REGISTERs: 3333 (100.0%) Failed re-REGISTERS: 0 (0.0%) REGISTER retransmissions: 0 Average time from REGISTER to 200 OK: 49.0ms ----------------------------- Statistics Screen ------- [1-9]: Change Screen -- Start Time | 2018-04-27 18:11:59.203632 1524832919.203632 Last Reset Time | 2018-04-27 18:13:06.570393 1524832986.570393 Current Time | 2018-04-27 18:13:06.570934 1524832986.570934 -------------------------+---------------------------+-------------------------- Counter Name | Periodic value | Cumulative value -------------------------+---------------------------+-------------------------- Elapsed Time | 00:00:00:000000 | 00:01:07:367000 Call Rate | 0.000 cps | 16.076 cps -------------------------+---------------------------+-------------------------- Incoming call created | 0 | 0 OutGoing call created | 0 | 1083 Total Call created | | 1083 Current Call | 0 | -------------------------+---------------------------+-------------------------- Successful call | 0 | 1083 Failed call | 0 | 0 -------------------------+---------------------------+-------------------------- Response Time 1 | 00:00:00:000000 | 00:00:00:119000 Call Length | 00:00:00:000000 | 00:00:07:413000 ------------------------------ Test Terminated -------------------------------- Another thing from Rob's answer: ->to increase the calls per second independently of registration rate, you’d need to edit the script. where and what I need to change? I tried to change in run_stress BHCA = 1.3 i increase it to 4.3 OUTGOING_BHCA = BHCA/2.0 make it BHCA/1.0 RE_REG_PER_HOUR = 2.0 make it to 1.0 Regards, Arun Lal On Thu, Apr 26, 2018 at 5:09 PM, William Yates <[email protected]<mailto:[email protected]>> wrote: Hi Arun, This subject has been raised recently, with some comprehensive answers from Rob Day here - please have a read: http://lists.projectclearwater.org/pipermail/clearwater_lists.projectclearwater.org/2018-April/003866.html More nodes seems like the right approach, but it's difficult to predict what you would need to achieve your aims. I hope this helps, but get back if not. Regarding the ZeroDivisionError, there is already an issue to cover this: https://github.com/Metaswitch/project-clearwater-issues/issues/30 Cheers, Will From: Clearwater [mailto:[email protected]<mailto:[email protected]>] On Behalf Of Arun Lal Sent: 24 April 2018 17:37 To: [email protected]<mailto:[email protected]> Subject: Re: [Project Clearwater] Regarding Increasing #calls using stress testing Hi guys, it give following result: []ubuntu@stress:~$ /usr/share/clearwater/bin/run_stress rags.mf 40000 1 --multiplier=12 Starting initial registration, will take 500 seconds Initial registration succeeded Starting test Test complete Elapsed time: 00:01:43 Start: 2018-04-25 05:28:52.781011 End: 2018-04-25 05:30:57.535772 Total calls: 5200 Successful calls: 5168 (99.3846153846%) Failed calls: 32 (0.615384615385%) Unfinished calls: 0 Retransmissions: 0 Average time from INVITE to 180 Ringing: 10387.0ms # of calls with 0-2ms from INVITE to 180 Ringing: 0 (0.0%) # of calls with 2-10ms from INVITE to 180 Ringing: 0 (0.0%) # of calls with 10-20ms from INVITE to 180 Ringing: 0 (0.0%) # of calls with 20-50ms from INVITE to 180 Ringing: 0 (0.0%) # of calls with 50-100ms from INVITE to 180 Ringing: 0 (0.0%) # of calls with 100-200ms from INVITE to 180 Ringing: 0 (0.0%) # of calls with 200-500ms from INVITE to 180 Ringing: 0 (0.0%) # of calls with 500-1000ms from INVITE to 180 Ringing: 0 (0.0%) # of calls with 1000-2000ms from INVITE to 180 Ringing: 3 (0.0576923076923%) # of calls with 2000+ms from INVITE to 180 Ringing: 5167 (99.3653846154%) Failed: call success rate 99.3846153846% is lower than target 100.0%! Total re-REGISTERs: 16000 Successful re-REGISTERs: 15559 (97.24375%) Failed re-REGISTERS: 441 (2.75625%) REGISTER retransmissions: 0 Average time from REGISTER to 200 OK: 5149.0ms Failed: re-registration success rate 97.24375% is lower than target 100.0%! Log files at /var/log/clearwater-sip-stress/2679_* some time it gives like: []ubuntu@stress:~$ /usr/share/clearwater/bin/run_stress iind.intel.com<http://iind.intel.com> 50000 1 --multiplier=8 Starting initial registration, will take 625 seconds Initial registration succeeded Starting test Test complete Traceback (most recent call last): File "/usr/share/clearwater/bin/run_stress", line 346, in <module> call_success_rate = 100 * float(row['SuccessfulCall(C)']) / float(row['TotalCallCreated']) ZeroDivisionError: float division by zero I want to make 1 million calls per second (is it possible?) because I have tried by varying the #subscribers and multiplier and other parameter but it is not increasing. Firsy I want it is able to make 20K calls per sec (or 50K or 1lack per sec). Every node is working fine, I don't know why I am not able to scale the performance or # calls . If anything need to change in script run_stress, please let me know where and what. your reply will highly appreciate. Regards, Arun Lal On Tue, Apr 24, 2018 at 2:28 PM, Arun Lal <[email protected]<mailto:[email protected]>> wrote: Hi, Nice work by clearwater team, I really appreciate your work. The document of clearwater is self explanatory, most of the task I did by reading that, but I have stuck at some points: > I have 3 sprout, 2 vellum, 2 dime, 1 bono, 1 homer, 1 ellis nodes. I want to > make 15000 calls/sec or more. I can increase more nodes if possible. > I am not able to make more calls, I tried []ubuntu@stress:~$ /usr/share/clearwater/bin/run_stress rags.mf 30000 1 --multiplier=20 but it gives around 3000 calls. > what exact command should I use to make more calls per sec. All the node are > working fine I have check that zoiper and other sip client working fine. > Can you please share exact command so that I can increase no. of calls per > sec. Or how many more nodes are required, I will add. Thanks, Arun Lal Singh _______________________________________________ Clearwater mailing list [email protected]<mailto:[email protected]> http://lists.projectclearwater.org/mailman/listinfo/clearwater_lists.projectclearwater.org
_______________________________________________ Clearwater mailing list [email protected] http://lists.projectclearwater.org/mailman/listinfo/clearwater_lists.projectclearwater.org
