On Thu, 13 Dec 2001, Brian Ashe wrote:
> Here's my best shot at explaining this. Hope it is clear.
>
> Those numbers you see in top/w/uptime are the average number of processes in
> the run queue for the last 1, 5, 15 minutes respectively. The higher the
> numbers the higher the load.
>
> The problem with this is that the load averages are dependant upon the
> system. So there isn't a clear way to say what is too much. You need to know
> your system well enough to know if you are reaching it's capacity. A
> powerful (MIPS/RISC/whatever) system might still have plenty of juice left
> with load averages of 50 or more. A P-100 would likely start dying (becoming
> unresponsive) at around 5-6. So a load of 1 on one system is not the same as
> a load of 1 on another (unless they are identical).
>
> If you can push your system to it's limit, and monitor top while doing so,
> note the point at which the system becomes unresponsive (or extremely poor)
> and you will know where you stand at all other times.
>
What's considered unresponsive? Right now, the biggest limiting factor,
at least on the 266 w/ 64MB RAM, is when it runs out of physical memory,
and has to swap to the disk. Then, it runs like a damn 486 it seems ;)
Depending on what I have going at any given moment, that can be just
switching from one desktop to another in XFCE (a very lightweight wm) to
change from Opera to pine. Granted, I think some of that was due to some
of the problems in the earlier 2.4.x kernels. I've had RH 7.1 w/ the
stock kernel (2.4.3), then the updated 2.4.9, then SuSE 7.3 w/ stock
(2.4.7), and then the unofficial SuSE updated kernel 2.4.16, which is
running like a dream and seems to mitigate a little bit of the problem of
swapping. The other two boxes, one a dual-boot P3-500 w/ 384MB desktop
and a P2-400 w/ 256MB server are both running (currently) stock RH 7.2. I
asked because I've seen load starting to get up to around 2 on these two
as well, though not for extended times, and not accompanied by the 'swap
storms' that the 266 had/has, but that's probably because it's a lot
harder to run out of RAM w/ 256 or 384MB than 64.
Just for shits and giggles, what would be a good way to stress test these
machines so I do have some idea what they can take, so as to lay my mind
to rest (if all goes well) about the current load averages?
TIA,
Monte
_________________________________________________________
Do You Yahoo!?
Get your free @yahoo.com address at http://mail.yahoo.com
_______________________________________________
Redhat-list mailing list
[EMAIL PROTECTED]
https://listman.redhat.com/mailman/listinfo/redhat-list
