On Sat, Jun 02, 2001 at 10:10:31AM -0700, A.J. Rossini wrote:
> >>>>> "v" == viral <[EMAIL PROTECTED]> writes:
>
> v> Let me know if you find out what happened. I have noticed heavy
> v> arp activity on my mosix cluster, when I was running 2.2.17 at
> v> times, which I couldn't explain for. Someone on this list said
> v> that he experienced the same thing if a mosix node went
> v> down. Do you see any such activity ?
>
> I do -- or if a node is "closed" for recieving.
Is there any solution to this ? Nodes can always go down.. I could cross-post
to the mosix mailing list, but I'm not on it as of now, and it won't allow
me to post..
> v> 1.0.3 debs are already in unstable. :)
>
> Thanks for the pointers for building the kernel in a previous email!
I'm glad if it helped !
> BTW, 2.4.[45] have nasty, nasty habits of not freeing swap[1]. I've
> gotten burned by this (some of the simulations I run are moderately
> big, pushing things into swap on a reasonably big machine).
>
> [1]: recent Kernel-Traffic report
And I was wondering, why my machine with 128MB RAM was using up 50 MB
swap consistently !
Is this going away anytime soon ? patches in the -ac series kernels ?
> you might consider the 2.2.19 mosix patches (not sure how to include
> them, something like mosix2.2.19 package (yech), or similar -- but
> I've got no better suggestions).
>
> 2.2.19 + Mosix 0.98 has been semi-rock solid, from what I've been
> observing...
Maybe we can have mosix, and mosix0.98 then ? I'll try to put this in
sometime next week.
viral
--
And if your head explodes with dark forebodings too,
I'll see you on the dark side of the moon.
PGP signature
