Extracting request elapsed time values using blkparse

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Dear all,

First, I would like to thank you all for your efforts.

I have been trying to use blkparse/blktrace to trace io activity for
my research.
I want to to measure IO response time to better characterize linux io.
My problem is that I am trying to use blkparse to get the request response time
for some activity.

I gave the -t switch along with %u to print it out.
While it does work for some requests. It gives "Expecting elapsed
value" and bails out. When I turn on verbose output, it also generates
"complete not found for XXYYZZ" when I request verbose results.

>From what I gather. Some blocks return completion twice for synchronous writes.
But it can also happens for reads.
I am attaching an annotated excerpt from the results. In these results
I have intentionally disabled the bailout accompanying "Expecting
elapsed value" so blkparse would not exit.

For a trace that has ~1.4M requests, I have seen this error 90K times.

The arguments are:
./blkparse -v -t -F C,"COMP %D %2c %8s %16S %5T.%9t %5p %2a %3d %3u\n"
-F D,"ISSUED %D %2c %8s %16S %5T.%9t %5p %2a %3d\n" -i
./results/sda.blktrace.

NOTE: This still happens well into the tracing process. E.g. Multiple
hours into the tracing process. But you should have commented the
exit(1) after fprintf(stderr, "Expecting elapsed value\n"); Otherwise,
it will bail out at the first occurance.

Any help will be greatly appreciated.
Regards,
-Tarihi

Attachment: broken-blkparse
Description: Binary data


[Index of Archives]     [Netdev]     [Linux Wireless]     [Kernel Newbies]     [Security]     [Linux for Hams]     [Netfilter]     [Bugtraq]     [Yosemite News]     [MIPS Linux]     [ARM Linux]     [Linux RAID]     [Linux Admin]     [Samba]

  Powered by Linux