On Fri, 2011-07-29 at 14:35 +0100, John Hodrien wrote:
On Fri, 29 Jul 2011, Jakub Hrozek wrote:
> On 07/29/2011 03:25 PM, John Hodrien wrote:
>>>> I'm thinking sssd likely to crawl into a hole at some point anyway,
but the
>>>> caching's a whole lot more sane, so I'm hoping it'll plod
along more
>>>> successfully than nss_ldap/nscd.
>> Just for kicks and giggles I pointed it at the entire tree (rather than the
>> normal subset I use), and ran 'id myusername'.
>>
>> First call:
>>
>> real 3m45.453s
>>
>> 182M cache_default.ldb
>
> This should be much faster once 1.6.0 is out (next week, hopefully),
> because we'll be using the ASQ dereferencing control[1] rather than
> sequentially requesting all group members.
>
> [1]
http://msdn.microsoft.com/en-us/library/aa746418%28v=vs.85%29.aspx
So if I tested again git head, that would be representative of what I might
expect from 1.6?
I'm more than happy to provide a data point for just how good you are against
a big installation!
Yeah, you can use git head, or drop
http://jdennis.fedorapeople.org/ipa-devel/ipa-devel-rhel.repo
into /etc/yum.repos.d and "yum update sssd" (that's our development
repository with all of the latest committed bits). Easier than building
it yourself :)