Check out this uptime
Discussion
Ok, for the non unix geeks this won't mean much, but just did an "uptime" on a slow box on our network and got this result:
2:00pm up 453 days, 16:19, 1 user, load average: 150.23, 147.34, 140.52
Load average 150!! And the box is still responding (just about - well enough to fix it).
For the non technically minded the load average on this box is usually around 0.3, and load is an indicator of the amount of jobs in the processor queue at any one time. For optimal performance this should usually be under 1.
2:00pm up 453 days, 16:19, 1 user, load average: 150.23, 147.34, 140.52
Load average 150!! And the box is still responding (just about - well enough to fix it).
For the non technically minded the load average on this box is usually around 0.3, and load is an indicator of the amount of jobs in the processor queue at any one time. For optimal performance this should usually be under 1.
_dobbo_ said:
I get all sniffy about my stats being ruined then realise that's a bit
its very nerdy but totally understandable. "power cycle that box for no good reason and i'll beat you to death with it". that threat usually worked for me in the past.
i know i had one box that was cruising through a high uptime, 400 odd days and going like a train. until some telehouse engineer for no reason at all mistook it and pulled its power. he was later buried in their gardens.
rebelstar said:
There is (was?) a bug where the load average could get really high but it was a false reading. I've seen it myself many times but I forgot what causes it. For some reason I keep thinking it's NFS related but I can't remember why that might be.
Processes that are hung waiting for filesystems add to the load average, but consume no other resources. It's not really a bug so much as people's misunderstanding of what the load average actually is; the number of processes that cannot run because they are waiting for a resource.
If an NFS server "goes away", all the processes using it on a client will immediately go onto the load average.
Mr E said:
500+ days? pah, mere amatuers this is from one of our solaris boxes. its an old SS20 with Dual Ross procs that just runs and runs.
> uptime
5:35pm up 1519 day(s), 20:54, 2 users, load average: 0.07, 0.07, 0.09
Not doing much though, is it?
not really. serves about 8,000 zone files, relays 100,000 emails a day for a few customers, and runs a few other boring things like backup authentication daemons and that sort of stuff. spent a lot of time making that box run cool. its only when my business partner puts some of the most awful perl scripts on does it start to get a bit stressed. i've never seen this one get past about 45 on load during the current uptime. also the Ross processors give it a fair turn of speed for an old lady.
The primary MX for my domain, which is an old Compaq with 32MB RAM and a P120 processor, sitting on a shelf in my study, running a self modified version of qmail on debian potato (2.2.17 kernel to give you an idea of how old it is)
mark@dangermouse:~$ uptime
15:21:08 up 138 days, 16:00, 1 user, load average: 0.00, 0.00, 0.00
The highest it ever got to was 700 days or so, but SWEB put a stop to that by dipping the power for 2 seconds one morning.
mark@dangermouse:~$ uptime
15:21:08 up 138 days, 16:00, 1 user, load average: 0.00, 0.00, 0.00
The highest it ever got to was 700 days or so, but SWEB put a stop to that by dipping the power for 2 seconds one morning.
Gassing Station | Computers, Gadgets & Stuff | Top of Page | What's New | My Stuff