cclogd and diagmond

[Ted Ellis_2]

I have upgraded online diags to the March 01 realease and am now seeing periodic cases on N4000 servers (11.00) with 1 cpu running 100% with cclogd and diagmond using up to 40% on another. Anyone know the cause of this "spin lock"?

[melvyn burnard]

This has been asnwered before, and appears to be a known issue with the revision you have loaded. It will be fixed in the next (June I believe) release.
Search on the forums for the cclogd and you should get some other discussions on this subject.

[Phil Squire]

We have exactly the same problem with our N-class server running HP-UX 11.00 but an earlier version of Diagnostics. HP support told us the problem would be fixed in the March 2001 release - wrongly it seems.

You can alleviate your server of it's burden by switch off the diagnostic subsystem if you need to:

/sbin/init.d/diagnostic stop

[William Baines]

Hi Ted,

I had the same problem since installing MAR2001 Online Diags.

HP just released an STM patch that includes a fix for cclogd spin-lock:

Symptoms:
PHSS_23655:
2. Chassis code logging daemon(cclogd) consumes all processor cycles during archiving of boot logs causing possible system hang.

[William Baines]

Ooops - just tried to load the patch PHSS_23655 and discovered it's for STM A.24 (B.11.00.15.08). I'm running STM A.25 (B.11.00.16.09). Guess we'll have to wait a bit longer for HP to patch STM A.25 for cclogd spin-lock. *sigh*