On Oct 19, 2010, at 11:22 AM, Bruce Momjian wrote:
> Greg Smith wrote:
>> A.M. wrote:
>>> Perhaps a simpler tool could run a basic fsyncs-per-second test and prompt the DBA to check that the numbers are
withinthe realm of possibility.
>>>
>>
>> This is what the test_fsync utility that already ships with the database
>> should be useful for. The way Bruce changed it to report numbers in
>> commits/second for 9.0 makes it a lot easier to use for this purpose
>> than it used to be. I think there's still some additional improvements
>> that could be made there, but it's a tricky test to run accurately. The
>
> test_fsync was designed to test various things like whether several
> open-sync writes are better than two write and an fsync, and whether you
> can fsync data written on a different file descriptor. It is really a
> catch-all test right now, not one specific for choosing sync methods.
I am working on simplifying the test_fsync tool and making it a contrib function which can be run by the superuser
basedon the configured fsync method. That way, the list can ask a user to run it to report fsyncs-per-second for
suspiciousness.The goal is to make it more accessible. I was also thinking about adding some notes along the lines of
"Yourdrive fsync speed rates between a 5400 RPM SATA drive and a 7200 RPM SATA drive." or "Your drive fsync speed rates
ashigh as RAM- your fsync method may be wrong."
Currently, the test tool is not even compiled by default.
Thoughts?
Cheers,
M
