On Mon, Jan 31, 2005 at 08:32:39PM +0000, Greg Kochanski wrote: > Horms wrote: > >On Sun, Jan 30, 2005 at 04:38:05PM +0000, Greg Kochanski wrote: > > > >>Package: kernel-image-2.6.8-2-686 > >>Version: 2.6.8-12 > >>Severity: important > >> > >> > >>Load average is reported very high, but it ain't so. > >> > >>$ cat /proc/loadavg > >>52.80 51.80 50.81 4/202 8431 > >>$ > >> > >>The system is currently unloaded and quite responsive. > >>The real load average is clearly less than 2, > >>and from looking at top, the load average is probably > >>about 0.1 . > > > > > >In Bug#29285 you report that you have processes hung while > >accessing the filesystem. I would strongly suspect that is related to > >this load. > > > Possibly, but those processes were showing zero CPU, > as displayed by "top". The system was also far too responsive > to be heavily loaded in reality.
load average is not a measure of how busy your CPU is. It is a measure of how many processes are waiting for system resources. Usually high CPU utilisation and high load go together, however they are not the same thing. I would syspect that the load average being reported it is indeed accurate, and that some resource is in contention. > >>This is breaking my mail; exim4 is configured on my > >>system to refrain from delivering or accepting mail > >>when the load average is too high. > > > > > >I suspect that this load average is a symtom of some breakage, > >not a cause. > > The load average is a direct cause of exim4's breakage, > because it was configured to stop accepting mail when > the load average was too high. > What caused the bad load average, I don't know. > > Previous to the high load average, I had been trying > to mount a SCSI disk over USB. This may also be related. Is dmesg showing you anything useful. -- Horms -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
