Skip to Content.
Sympa Menu

perfsonar-user - Re: [perfsonar-user] PSCHEDULER BUG: Pscheduler maximum limit amount of time to schedule a task

Subject: perfSONAR User Q&A and Other Discussion

List archive

Re: [perfsonar-user] PSCHEDULER BUG: Pscheduler maximum limit amount of time to schedule a task


Chronological Thread 
    < Chronological >      < Thread >
  • From: Murilo Vetter <>
  • To: Mark Feit <>
  • Cc: perfsonar-user <>, monipe-des <>
  • Subject: Re: [perfsonar-user] PSCHEDULER BUG: Pscheduler maximum limit amount of time to schedule a task
  • Date: Thu, 5 Dec 2019 12:13:39 -0300 (BRT)
Hi Mark

To exemplify better what I was saying that it does not show server site test, I send the file attached.

Hope it helps to identify the problem.

Regards,
--
Murilo Vetter
Analista de Tecnologia da Informação
REMEP-FLN - Rede Metropolitana Comunitária de Educação e Pesquisa da Região de Florianópolis
PoP-SC - Ponto de Presença da RNP em Santa Catarina
RNP - Rede Nacional de Ensino e Pesquisa
+55 (48) 3721-3000
www.pop-sc.rnp.br | www.rnp.br

De: "Murilo Vetter" <>
Para: "Mark Feit" <>
Cc: "perfsonar-user" <>, "monipe-des" <>
Enviadas: Quinta-feira, 5 de dezembro de 2019 8:36:54
Assunto: Re: [perfsonar-user] PSCHEDULER BUG: Pscheduler maximum limit amount of time to schedule a task

Hi Mark,

One thing that I was thinking about is that in our scenario, we are deleting file: /etc/pscheduler/limits.conf

The purpose of that was to bypass some limitations imposed by default restrictions.

We would like to run long time period tests. In the past, we already run this kind of test invoking directly iperf3 tool.

Do you think could it interfere in anything? If pscheduler does not find the limit file, are there default values it consider?

Bests,
--
Murilo Vetter
Analista de Tecnologia da Informação
REMEP-FLN - Rede Metropolitana Comunitária de Educação e Pesquisa da Região de Florianópolis
PoP-SC - Ponto de Presença da RNP em Santa Catarina
RNP - Rede Nacional de Ensino e Pesquisa
+55 (48) 3721-3000
www.pop-sc.rnp.br | www.rnp.br

De: "Murilo Vetter" <>
Para: "Mark Feit" <>
Cc: "perfsonar-user" <>, "monipe-des" <>
Enviadas: Quarta-feira, 4 de dezembro de 2019 12:53:37
Assunto: Re: [perfsonar-user] PSCHEDULER BUG: Pscheduler maximum limit amount of time to schedule a task

Hi Mark,

As I could understood, long time tests in versions 4.2.* it is now outputing the results.

And in the versions 4.1.4 and 4.1 it is writing the results from client side, but the server side is not report anything.

Version 4.1 and 4.1.4, when trying to run:


The server side outputs:

Diagnostics from 10.39.2.6:
    (No diagnostics)

Error from 10.39.2.6:
    iperf3 returned an error: 
    Process took too long to run.

The schedule tests:

[root@470fce348303 /]# pscheduler schedule -PT1H
2019-12-03T10:17:01-05:00 - 2019-12-04T10:07:03-05:00  (Finished)
rtt --count 85800 --dest 10.39.2.6 --source 10.39.2.2 (Run with tool 'ping')
https://localhost/pscheduler/tasks/a4f8ddd9-0517-4ad3-961f-cbeb2d9105df/runs/c1b4228d-c218-4bfc-9280-1fc4d27416fa


2019-12-03T10:17:04-05:00 - 2019-12-04T10:07:11-05:00  (Finished)
throughput --bandwidth 100000000 --duration PT23H50M --source 10.39.2.2 --dest 10.39.2.6 --interval PT1M --buffer-length 1472 --udp (Run with tool 'iperf3')
https://localhost/pscheduler/tasks/43b8933b-a910-4ed3-a031-73a315f094f5/runs/9c7a523f-2b40-4aa0-8d6d-ff927ccc0ac7

Is something that I am missing?

Regards,
--
Murilo Vetter
Analista de Tecnologia da Informação
REMEP-FLN - Rede Metropolitana Comunitária de Educação e Pesquisa da Região de Florianópolis
PoP-SC - Ponto de Presença da RNP em Santa Catarina
RNP - Rede Nacional de Ensino e Pesquisa
+55 (48) 3721-3000
www.pop-sc.rnp.br | www.rnp.br

De: "Mark Feit" <>
Para: "Murilo Vetter" <>
Cc: "perfsonar-user" <>, "monipe-des" <>
Enviadas: Segunda-feira, 2 de dezembro de 2019 12:55:12
Assunto: Re: [perfsonar-user] PSCHEDULER BUG: Pscheduler maximum limit amount of time to schedule a task

Murilo Vetter writes:

 

Hi Mark,

 

We already tried to run this on perfsonar docker images:

- Latest (4.2.2)

- v4.2.1

 

Now I am going to try to run on 4.1.4 to see if it is going to have problem.

Do you know if it is a problem on newer version 4.2.*?

 

The code that covers this hasn’t changed since it was first written, so I would expect the behavior to be the same on all versions.  We’re planning a bugfix release in the next few weeks; whatever fix I come up with will likely ship with that.  I’ve been up to my neck in SC for most of the last month and now am up to my neck in TechEx, so your patience is appreciated.

 

--Mark

 


docker exec monipe-src-b23d3198-bcc9-45de-a0c3-4e30ee20482a pscheduler result
--diags
https://10.39.2.2/pscheduler/tasks/5573d495-ab08-4444-b0bc-25c51858c870/runs/d8c88348-aa69-4f9a-8b23-280f7612afab

2019-12-02T08:17:04-05:00 on 10.39.2.2 and 10.39.2.6 with iperf3:

throughput --bandwidth 100000000 --duration PT23H50M --source 10.39.2.2 --dest
10.39.2.6 --interval PT1M --buffer-length 1472 --udp

* Stream ID 5
Interval Throughput Lost / Sent
0.0 - 60.0 100.00 Mbps 509504 sent
60.0 - 120.0 100.00 Mbps 509511 sent
120.0 - 180.0 100.00 Mbps 509511 sent
180.0 - 240.0 100.00 Mbps 509511 sent
240.0 - 300.0 100.00 Mbps 509511 sent
300.0 - 360.0 100.00 Mbps 509511 sent
360.0 - 420.0 100.00 Mbps 509510 sent

.. OMMITED ..

85560.0 - 85620.0100.00 Mbps 509511 sent
85620.0 - 85680.0100.00 Mbps 509511 sent
85680.0 - 85740.0100.00 Mbps 509511 sent
85740.0 - 85800.099.96 Mbps 509511 sent

Summary
Interval Throughput Lost / Sent
0.0 - 85800.1 100.00 Mbps 14964660 / 728600537Jitter: 0.114149184212 ms

Limit system diagnostics:
No limits were applied


Diagnostics from 10.39.2.2:
/usr/local/src/iperf3.sh -p 5201 -B 10.39.2.2 -c 10.39.2.6 -t 85800
--json -i 60 -b 100000000 -u -l 1472

{
"start": {
"connected": [{
"socket": 5,
"local_host": "10.39.2.2",
"local_port": 58005,
"remote_host": "10.39.2.6",
"remote_port": 5201
}],
"version": "iperf 3.6",
"system_info": "Linux 591912125ab3 3.10.0-957.5.1.el7.x86_64
#1 SMP Fri Feb 1 14:54:57 UTC 2019 x86_64",
"timestamp": {
"time": "Mon, 02 Dec 2019 13:17:07 GMT",
"timesecs": 1575292627
},
"connecting_to": {
"host": "10.39.2.6",
"port": 5201
},
"cookie": "g3opxcgrlxqmxxfcotlej2ynylpu7ye6p7zu",
"sock_bufsize": 0,
"sndbuf_actual": 212992,
"rcvbuf_actual": 212992,
"test_start": {
"protocol": "UDP",
"num_streams": 1,
"blksize": 1472,
"omit": 0,
"duration": 85800,
"bytes": 0,
"blocks": 0,
"reverse": 0,
"tos": 0
}
},
"intervals": [{
"streams": [{
"socket": 5,
"start": 0,
"end": 60.000077962875366,
"seconds": 60.000076293945312,
"bytes": 749989888,
"bits_per_second":
99998524.5786339,
"packets": 509504,
"omitted": false
}],
"sum": {
"start": 0,
"end": 60.000077962875366,
"seconds": 60.000076293945312,
"bytes": 749989888,
"bits_per_second": 99998524.5786339,
"packets": 509504,
"omitted": false
}
},


.. OMMITED ..



"cpu_utilization_percent": {
"host_total": 10.3519663168971,
"host_user": 1.2856917410920921,
"host_system": 9.066287776237175,
"remote_total": 13.533644117372637,
"remote_user": 2.6556851499717258,
"remote_system": 10.877965729380499
}
}
}


Diagnostics from 10.39.2.6:
No result was produced

Error from 10.39.2.6:
No result was produced

Archive powered by MHonArc 2.6.19.

Top of Page