StoRM at CNAF

  • Thu Oct 4 2007

Analysis of run storm-fe.test7 from Thu Oct 4 19:16 to Thu Oct 4 22:13 2007
Num of parallel proc.: 100 each with 100 requests
Increasing polling time, starting with t0=8, timeout for a request: 8176.0 s (136.3 min.)
Mean time per request in each process: 62.8 s ( 1.0 min.)
Average over all the run: (total req. done)/(total duration)= 0.939 s (frequency= 1.06 Hz)
Total failures getting the request token: 0
Total failures getting the TURL: 41
Total failures in ptp => 41 over 10000 (0.41 %)
Comments: 40 out of 41 failed requests are due to a pure timeout (no error occurred.) Then one error occurred to get a TURL: the request token is correctely returned and the following polling show that the request status is SRM_REQUEST_INPROGRESS. But then, at the fifth polling the error message is displayed (at time: Thu Oct 4 20:58:28 2007):

Sending StatusPtP request to: httpg://storm-fe.cr.cnaf.infn.it:8444/
 ============================================================
Request status:
  statusCode="SRM_FAILURE"(1)
statusptp:10:   explanation="Generic error quering the status for the request"

  • Tue Oct 2 2007

Analysis of run storm-fe.test6 from Tue Oct 2 16:38:22 2007 to Tue Oct 2 18:56:11 2007
Num of parallel proc.: 100 each with 100 requests
Increasing polling time, starting with t0=4, Max. timeout for a request: 4088.0 s ( 68.1 min.)
Mean time per request in each process: 52.0 s ( 0.9 min.)
Average over all the run: (total req. done)/(total duration)= 1.209 s (frequency= 0.83 Hz)
Total failures getting the TURL: 40
Total failures in ptp => 40 over 10000 (0.4 %)
Comments All failed request are due to the timeout implemented in the client script. Run again the same test with a less restrictive timeout.

Analysis of run storm-fe.test5 from Tue Oct 2 14:10:07 2007 to Tue Oct 2 16:20:43 2007
Num of parallel proc.: 100 each with 100 requests
Increasing polling time, starting with t0=2, Max. timeout for a request: 2044.0 s ( 34.1 min.)
Mean time per request in each process: 55.2 s ( 0.9 min.)
Average over all the run: (total req. done)/(total duration)= 1.276 s (frequency= 0.78 Hz)
Total failures getting the TURL: 90
Total failures in ptp => 90 over 10000 (0.9 %)
Comments The failed requests are due to the timeout implemented in the client script. Run again the same test with a less restrictive timeout.

Analysis of run storm-fe.test4 from Tue Oct 2 10:43:33 2007 to Tue Oct 2 12:23:08 2007
Num of parallel proc.: 100 each with 100 requests
Increasing polling time, starting with t0=2, Max. timeout for a request: 2044.0 s ( 34.1 min.)
Mean time per request in each process: 40.2 s ( 0.7 min.)
Average over all the run: (total req. done)/(total duration)= 1.674 s (frequency= 0.60 Hz)
Total failures getting the request token: 0
Total failures getting the TURL: 4503
Total failures due to SRM_INVALID_PATH: 4428
Total failures due to SRM_DUPLICATION_ERROR: 0
Total failures in ptp => 4503 over 10000 (45.03 %)
Comments Again problems of network at CNAF. Results do not make sense. The run has to be repeated.

  • Fri Sep 28 2007

Analysis of run storm-fe.test3 from Thu Sep 27 17:36:51 2007 to Thu Sep 27 19:42:38 2007
Num of parallel proc.: 100 each with 100 requests
Increasing polling time, starting with t0=4
Max. timeout for a request: 4088.0 s ( 68.1 min.)
Mean time per request in each process: 66.1 s ( 1.1 min.)
Average over all the run: (total req. done)/(total duration)= 1.325 s (frequency= 0.75 Hz)
Total failures in ptp => 0 over 10000 (0 %)
Comments: the UI used to run the client script (ui01-lcg.cnaf.infn.it) is having problem (it is extremely slow). The origin of the problem is the GPFS file system, which is affecting CNAF facilities.

  • Thu 27 Sep 2007
Analysis of run storm-fe.test1 from Thu Sep 27 14:23:46 2007 to Thu Sep 27 14:34:30 2007
Num of parallel proc.: 10 each with 100 requests
Increasing polling time, starting with t0=2, Max. timeout for a request: 2044.0 s ( 34.1 min.)
Mean time per request in each process: 5.9 s ( 0.1 min.)
Average over all the run: (total req. done)/(total duration)= 1.553 s (frequency= 0.64 Hz)
Total failures in ptp => 0 over 1000 (0 %)

Analysis of run storm-fe.test2 from Thu Sep 27 15:22:54 2007 to Thu Sep 27 16:58:31 2007
Num of parallel proc.: 100 each with 100 requests
Increasing polling time, starting with t0=2 Max. timeout for a request: 2044.0 s ( 34.1 min.)
Mean time per request in each process: 33.9 s ( 0.6 min.)
Average over all the run: (total req. done)/(total duration)= 1.743 s (frequency= 0.57 Hz)
Total failures getting the request token: 0
Total failures getting the TURL: 2442
Total failures due to SRM_INVALID_PATH: 2379
Total failures due to SRM_DUPLICATION_ERROR: 0
Total failures in ptp => 2442 over 10000 (24.42 %)
Comments: the GPFS cluster was down during 2 hours. Repeat the same test with same parameters.

-- ElisaLanciotti - 27 Sep 2007

Edit | Attach | Watch | Print version | History: r9 < r8 < r7 < r6 < r5 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r9 - 2007-10-05 - ElisaLanciotti
 
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    LCG All webs login

This site is powered by the TWiki collaboration platform Powered by PerlCopyright & 2008-2020 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback