[SOLVED] Server Load

[davidfsmith]

so 500 Users on a RHEL 4.0 64 bit install (12Gb RAM 1Tb HDD, Dual XEON), under 4.5.10 load averages never went above 2.0

Day two of the working year and of 5.0 being used more in anger and I'm looking at top:

top - 09:48:03 up 6 days, 21:47,  1 user,  load average: 35.12, 37.74, 38.37
Tasks: 308 total,   4 running, 304 sleeping,   0 stopped,   0 zombieCpu(s): 91.0% us,  7.3% sy,  0.0% ni,  0.0% id,  0.0% wa,  0.4% hi,  1.3% si
Mem:  12302208k total, 11972836k used,   329372k free,   383768k buffers
Swap:  8388600k total,      208k used,  8388392k free,  4342436k cached

So of course we're getting email delays. (80 minutes so far on internal email)

any one got some optimisation tips ?

[davidfsmith]

Ok the reasons for the load shown by top are seemingly nothing to do with Zimbra and a lot more to do with a server elsewhere trying to spam our server to death, this has now been resolved and with it the server loads are back to more normal levels....

*phew*

Now to get back to trying to get IM working

[SpecialKdkj]

Congratulations on achieving the highest Linux load averages I've ever seen. I once broke 15 and thought that was high!

[Klug]

I went above 50 this week-end (not on a Zimbra server however : 40 qmail-smtpd threads in parallel, each one calculating its own 1024 DH key, on a P3-800 server).

Marking this thread as solved.

[davidfsmith]

I went above 50 this week-end (not on a Zimbra server however : 40 qmail-smtpd threads in parallel, each one calculating its own 1024 DH key, on a P3-800 server).

Marking this thread as solved.

Hit over 40 later in the morning which was nice....

;-)

[p24t]

I've seen over 300 on a web server. It might've gone higher, but we'll never know because it stopped responding.

[mlombardi]

how do you manage to resolve the issue??

I have a similar issue and I dont know where to look.

Regards,
Max