Anthony Brock wrote:
> > I have 4 GB of physical RAM in the Ubuntu box + 3 GB swap. It is
> > running a vanilla 126.96.36.199 kernel (no skas3 patches, because I haven't
> > successfully managed to apply any sets of skas3 patches)
> Here is an interesting problem. Each instance has 48M. You're launching 50
> instances. This means you're allocating up to 2.4 GB for instances. Howev=
> you're also giving them 3 GB of swap. Do they need the swap? If not, get =
> of it (or at least lock your existing memory into ram).
If you are using sparse file for swap files I don't think it will take more=
than tenth of that space.
> > guest machines randomly crash. They aren't sending me anything via
> > syslog (I've configured them to log to the host machine), and they don't
> > appear to log anything to STDOUT. I can make machines crash with the
> > message in the subject line, by simply working them too hard (e.g. log
> > in, fire off a few dozen shells, add about 1000 users as fast as bash
> > can go through a for loop). My students, however (up to 25 working
> > simultaneously on UML guests), are making the machines crash by doing
> > nothing more than adding two or three users manually, looking at man
> > pages, and typing 'ls'.
> 50 machines doing ls (including swapping memory to and fro) could cause
> enough IO cause significant pauses. However, I can't explain the crashes.
> UML has been quite stable for me even under heavy load.
I had a problem where umls could sometimes crash and I found out that my /d=
ev/shm was too little. After I set it much bigger I haven't had any problem=
s with stability.
From: Jeff Dike <jdike@ad...> - 2006-09-15 16:07:59
On Mon, Sep 04, 2006 at 04:58:08PM +0300, jjkola@... wrote:
> If you are using sparse file for swap files I don't think it will
take more than tenth of that space.
Swapping onto a sparse file is really a bad idea. What happens if the
host filesystem fills up and the UML tries to swap? The swap writeout
will get an EIO, and, at best, the process being swapped will be killed.
> I had a problem where umls could sometimes crash and I found out
> that my /dev/shm was too little. After I set it much bigger I
> haven't had any problems with stability.
This results in a distinctive crash ("kernel mode signal 7" or
something similar), so this is unlikely to be the problem here.