Silkworm 12000 - Acceptable Load for a Full Switch?

[Alzhy]

Am not really that well versed yet with SAN Switches. Am somewhat versed as a UNIX/Linux SA and am just studying my way through FabricOS.

I noticed it is Linux based. So out of curiosity, I issued "uptime" and got the results below. The "load" immediately alarmed me. Normally for a Linux server - such load averages is already alarming.

Would this be normal for a SW 12000 which is fully populated?



fabric01:admin> uptime
6:48pm up 15 days, 36 min, 2 users, load average: 19.88, 20.04, 20.14


Thanks.

[Hein van den Heuvel]

Hi Nelson, looks like you mis-posted this reply in the "Digital image" familly instead of "storage". I'll ask a 'move'.

As for the problem itself, yeah that output would worry me also, but the switch is working fine right? Or at least reasonable?

Check with other tools? top, vmstat,...

This could be 'bad timing' with lots of threats runnable at teh sample times, but each action only taking microsecond (fibre interupts).

fwiw,
Hein.

[Patrick Terlisten]

Hello Nelson,

is the switch running fine, or are you facing some problems with it?

What is load? Load is the number of jobs with status R or D in the run queue.

So if you dont't have any problems, ignore the load. :) Linux is AFAIK only the underlaying OS. In fact the main work is done by a software called "FabOS". You can see this if you upgrade a Brocade Switch. There are some RPM packages called "FabOS".

Hope this helps.

Best regards,
Patrick