smbd 2.2.0; Solaris; performance woes?

[Jeremy Allison]

David Lee wrote:
> 
> A follow-up to my own question.  Might this load be due to come contention
> of these ~600 smbd processes onto the various ".tdb" files (and that
> general mechanism) introduced at 2.2.0?

Only is truss shows significant time in fcntl locks. Which
I doubt. tdb's are multi-reader/multi-writer (which is why
they're so fast).

We need more data to track this. It's more likely to be
a lpq cache time bug causing more lpq processes to be
spawned (a bug like that was fixed a few months ago in the
2.2 tree - but was being tested on Linux, not Solaris).

Jeremy.

-- 
--------------------------------------------------------
Buying an operating system without source is like buying
a self-assembly Space Shuttle with no instructions.
--------------------------------------------------------