Jerome,
The linux kernels are developed such that and 2.[even number] is stable
and any 2.[odd number] should be avoided, unless you like torture. See the
2.[odd number] stuff is development. The 2.[even number] is stable/release.
Also look for 2.[even number].[largest number] to get the newest stable
version. get 2.[odd number].[largest number] to get the newest craziest
wacked out kernel.
/B
----- Original Message -----
From: "Jérôme Bolliet" <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Thursday, January 09, 2003 23:56
Subject: Re: problem load-average with kernel 2.4.18-19.7.x
> Hi,
>
> Yes i know this.
> But my problem is : just by upgrading kernel, the memory usage and the
> load average explode.
>
> With 2.4.18-10, i'm using 700Mo of memory, and load is 1.
> With 2.5.18-19, i'm using 2 Go of memory and load is 3-4.
> With the same software and the same number of process !
>
> And i don't know why.
>
> Yoink! wrote:
> > On Thu, 9 Jan 2003, [ISO-8859-1] J?r?me Bolliet wrote:
> >
> >>We are using redhat 7.3 on Compaq DL360 2 processors with kernel
> >>2.4.18-10smp without problem to make POP3 server, with Courier IMAP
> >>1.6.2 and NetApp Filer F760 as NFS server.
> >>
> >>We have upgrade one server with latest kernel without changing
> >>parameter. And the load average are high.
> >>
> >>Top give us these information:
> >>
> >>2.4.18-10:
> >>
> >> 4:51pm up 2:49, 1 user, load average: 1.02, 1.14, 1.08
> >>113 processes: 109 sleeping, 4 running, 0 zombie, 0 stopped
> >>CPU0 states: 5.1% user, 4.3% system, 0.1% nice, 89.3% idle
> >>CPU1 states: 4.1% user, 6.4% system, 0.0% nice, 88.3% idle
> >>Mem: 2065192K av, 717504K used, 1347688K free, 0K shrd, 77308K
> >>buff
> >>Swap: 2097112K av, 0K used, 2097112K free 502364K
> >>cached
> >>
> >>2.4.18-19.7.x
> >>
> >> 4:51pm up 2 days, 2:00, 1 user, load average: 3.77, 3.07, 4.76
> >>107 processes: 106 sleeping, 1 running, 0 zombie, 0 stopped
> >>CPU0 states: 5.4% user, 7.0% system, 0.0% nice, 86.4% idle
> >>CPU1 states: 5.3% user, 8.0% system, 0.0% nice, 86.0% idle
> >>Mem: 2064880K av, 2047500K used, 17380K free, 0K shrd, 87404K
> >>buff
> >>Swap: 2097112K av, 0K used, 2097112K free 1782876K
> >>cached
> >>
> >>We can see that the load is 3 time higher and the memory is heavy used
> >>on the 19.7.x kernel.
> >
> >
> > You need to do a top and hit "M" to see which processes are using the
most
> > memory. If they aren't necessary, disable their startup.
> >
> > Also look at "vmstat 5 10" output and see if the "so" column reliably
> > stays above 0. If it does, your machine is swapping, and needs more
> > memory.
> >
> > You can _never_ look at used memory/swap as a sign of performance
trouble,
> > btw. Memory, and swap usage, no matter how much memory you have, should
> > approach ~98% usage and almost _never_ go back down. Only active
swapping
> > indicates a memory shortage.
> >
>
>
>
> --
> redhat-list mailing list
> unsubscribe mailto:[EMAIL PROTECTED]?subject=unsubscribe
> https://listman.redhat.com/mailman/listinfo/redhat-list
--
redhat-list mailing list
unsubscribe mailto:[EMAIL PROTECTED]?subject=unsubscribe
https://listman.redhat.com/mailman/listinfo/redhat-list
