[LU-1023] strange numbers shown in obdfilter-survey

• Type: Bug
• Resolution: Fixed
• Priority: Blocker
• Fix Version/s: Lustre 2.2.0
• Affects Version/s: Lustre 2.2.0
• Severity: 3
• Assignee: Di Wang
• Reporter: Shuichi Ihara (Inactive)

Description

I'm testing lustre-2.1.54 and quick ran obdfilter-survey on it. The obdfilter-survey could run, but some numbers are strange and some numbers are not working well. (same results on current master branch)

Here is obdfilter_survey_<date>_<hostname>.summary.

Mon Jan 23 12:07:10 JST 2012 Obdfilter-survey for case=disk from dl02
ost 1 sz 100663296K rsz 1024K obj 2 thr 128 write 714.81 ERROR read 1319.86 ERROR

It always shows ERROR for min and max value, because stats[0] in global values in obdfilter_survey_<date>_<hostname>.detail is always -1.

And, here is obdfilter_survey_<date>_<hostname>.detail.

126/128 Total: 637.928552/second
126/128 Total: 662.950942/second
126/128 Total: 564.928819/second
126/128 Total: 698.904949/second
128/128 Total: 696.935185/second
119/128 Total: 686.928559/second
Total: total 98304 threads 128 sec 0.000000 inf/second
=============> write global
-1 564.928819 828.885614

difftime (start and end) is zero(0.000000) and return infinity of thread per second, because divided by zero.

Also, I saw some very strange numbers on the detail log file.

128/128 Total: 805.908932/second
124/128 Total: 827.918036/second
128/128 Total: 851.893513/second
128/128 Total: 3098.609575/second
^^^^^^^^^^^^^^^^^^
128/128 Total: 798.896942/second

during the test, it sometimes shows unreasonable numbers above.

Issue Links

• Trackbacks: Changelog 2.2 version 2.2.0 Support for networks: o2iblnd OFED 1.5.4 Server support for kernels: 2.6.32220.4.2.el6 (RHEL6) Client support for unpatched kernels: 2.6.18274.18.1.el5 (RHEL5) 2.6.32220.4.2.el6 (RHEL6) 2.6.32.360....