On Thu, Feb 02, 2017 at 01:43:55PM +0100, Lukas Slebodnik wrote:
On (02/02/17 11:36), Jakub Hrozek wrote:
>On Wed, Jan 11, 2017 at 06:52:32PM +0100, Lukas Slebodnik wrote:
>> On (11/01/17 16:31), Jakub Hrozek wrote:
>> > *
https://fedorahosted.org/sssd/ticket/3074 - Move timestamp cache to
tmpfs
>> > *
https://fedorahosted.org/sssd/ticket/3097 - Measure the difference
between tmpfs database and NOSYNC database
>> > - tmpfs provides very little benefit, close with an explanation
>> Could you provide numbers/data?
>> It would be good to know what is a "very little benefit"
>> But I'm fine with closing #3074
>
>of course you are right. I re-run the tests and added measurements into
>#3097 (now I think 3097 can be closed because the purpose was only to
>add the measurements) and with what the benchmarks show is quite
>conslusive in the sense that #3074 can be closed or moved into patches
>welcome.
>
>Please let me know if you disagree, otherwise I'll close 3097 and move
>3074 into patches welcome.
I assume you tried to run it more time and you got simillar results
of course :) all the runs for this user were between 30-35 seconds for
both cases (the user was a member of about 500 groups)
I think we can close them.
OK