Hi List,
At the moment we are testing an NFS server based on a Xeon Nocona (em64t) and
2.6.11-rc3 compiled for X86_64/em64t as the machine has 8GB memory.
Filesystem is based on XFS with external journal.
This is a top snip when the server is under heavy read load:
<TOP>
top - 23:02:25 up 5:14, 6 users, load average: 8.54, 9.51, 8.99
Tasks: 122 total, 6 running, 116 sleeping, 0 stopped, 0 zombie
Cpu(s): 0.3% us, 92.4% sy, 0.0% ni, 0.0% id, 5.0% wa, 0.0% hi, 2.3% si
Mem: 8226600k total, 6946504k used, 1280096k free, 4212k buffers
Swap: 0k total, 0k used, 0k free, 6203480k cached
</TOP>
As you can see, the systime is ~90%.
I saw an earlier mail from Nathan regarding xfs_iget_core optimizations in
december, but as the following kernel profile shows, this procedure is
responsible for most of the load:
This is a kernel profile at the same time:
<PROFILE SNIP>
99621 total 0,0387
56322 xfs_iget_core 45,1298
33711 mwait_idle 234,1042
4500 xfs_iextract 10,8173
1077 __do_softirq 5,6094
284 tg3_start_xmit 0,1830
272 tg3_rx 0,2464
201 *unknown*
199 tg3_interrupt 0,5923
</PROFILE SNIP>
- Is this expected behavior? - or is something wrong (I would like to think
so :-))
- If you need any additional info I will be happy to provide it!
NOTE: This behavior does not show unless most of the memory is used for cache
(Right after reboot, systime is ~5%).
NOTE2: When only serving files which is already in cache systime also drops to
~5-10%
--
Med venlig hilsen - Best regards - Meilleures salutations
Anders Saaby
Systems Engineer
------------------------------------------------
Cohaesio A/S - Maglebjergvej 5D - DK-2800 Lyngby
Phone: +45 45 880 888 - Fax: +45 45 880 777
Mail: as@xxxxxxxxxxxx - http://www.cohaesio.com
------------------------------------------------
|