Hi Kishore - If could help if you could provide some performance data, since
you're talking about a performance problem.
Specifically, since you're doing large file reads (which I'll assume are
sequential
reads), what read throughput are you getting?
Thinking in terms of altering the memory page size for kernel heap segments
is a bit of stretch when you're looking to optimize ZFS sequential read
performance. And, again, not knowing what kind of performance you're actually
getting makes this difficult.
I'm not aware of any easy way to determine the page size used for kernel
heap segments. I need to look at the code and give that one a think.
You can measure read throughput at the application level using dtrace on
read:return:
#!/usr/sbin/dtrace -s
#pragma D option quiet
syscall::read:entry
/pid == $target && fds[arg0].fi_fs == "zfs"/
{
self->flag = 1;
}
syscall::read:return
/self->flag && errno == 0 /
{
@read_bytes = sum(arg0);
self->flag = 0;
}
tick-1sec
{
printa("bytes read per second: %...@d\n",@read_bytes);
trunc(@read_bytes);
}
Note I just typed in the script above - it looks OK but I did not actually
run it to test it. Note also the predicate on the entry probe.
Save the script as "rd.d", and run as;
rd.d -p <PID_OF_TARGET_PROCESS>
or
rd.d -c ./PROGRAM_NAME
Thanks,
/jim
Thanks,
/jim
On Aug 13, 2010, at 11:31 PM, Kishore Kumar Pusukuri wrote:
> Hi,
>
> One of my applications is spending around 90% of total execution time reading
> a huge file using read system call. I though that I could improve the
> performance of the application by increasing the page size for kernel heap. I
> know that I can increase page size of application heap using ppgsz command on
> the fly. But I don't know how to change the page size for kernel heap. Could
> anyone tell me if there is any command for this or is it possible through
> kdb/adb, please?
>
> My machine is a multi-core AMD Opteron running OpenSolaris.2009.06. File
> system is ZFS.
>
> And also please let me know if there are any ideas (tunable parameters) to
> improve the File IO on ZFS.
>
> Thank you.
>
> Best,
> Kishore
> --
> This message posted from opensolaris.org
> _______________________________________________
> perf-discuss mailing list
> [email protected]
_______________________________________________
perf-discuss mailing list
[email protected]
