Re: __alloc_pages: 2-order allocation failed
David, I'm not trying to be a pain, but I *am* trying to figure out
what's going on here. I've never seen these memory allocation errors
before, so while I can appreciate that spamc is likely not doing the
right thing, I'm wondering why the allocation errors are occuring in
the first place. Again, I've noted the following processes (including
the swapper!?) which have set off 1-order or 2-order allocation
complaints in the syslog.
Can you give me some idea as to why:
- All of these processes have caused this error
- This is the only box I've seen the errors on
- A box with nearly 6.5 GB of available memory (4.0 GB physical, ~ 2.5
GB swap), which mainly serves mail and websites for a university
department, is failing to allocate memory? I'm pretty positive that
it's not running out of memory to allocate...
Is it all related to spamc's failure to check errors on sockets? spamc
does seem to appear in (nearly?) every case - but if all of this is
happening just because spamc doesn't check for errors correctly, I'd
have to say that this is a major bug somewhere, as one errant process
shouldn't be causing problems like this.
On Thu, May 19, 2005 at 02:48:39PM -0400, Mike Edwards said:
> Is it possible that every single process listed here has the same
> issue?
>
> 1 task(courierpop3d)
> 1 task(dbclean)
> 1 task(qmail-remote)
> 1 task(qmail-to-mailma)
> 1 task(sh)
> 1 task(soffice.bin)
> 1 task(sperl5.8.4)
> 2 task(apache2)
> 2 task(smbd)
> 3 task(in.ftpd)
> 3 task(lpd)
> 3 task(top)
> 4 task(imapd)
> 4 task(python)
> 10 task(display)
> 10 task(spamd)
> 12 task(perl5.8.4)
> 26 task(abiword)
> 28 task(sshd)
> 541 task(spamc)
> 849 task(swapper)
>
> This is a list of processes that have set off this complaint, and how
> many times they've been seen (I'm rather disturbed that swapper has
> surpassed spamc...).
>
> It may very well be that spamc has a bug - but why am I seeing this
> error for all of these processes?
>
>
> On Thu, May 19, 2005 at 11:37:43AM -0700, David S. Miller said:
> > From: Mike Edwards <sauron-debian-sparc@psychology.rutgers.edu>
> > Date: Thu, 19 May 2005 13:25:17 -0400
> >
> > > David, do you know if anything further was done on this issue?
> >
> > The spamc daemons simply don't do correct error checking on
> > writes to AF_UNIX sockets. I thought I explained this pretty
> > well, so that someone could go check out the userspace bits
> > and look for the bug.
> >
--
Mike Edwards <sauron-debian-sparc@psychology.rutgers.edu>
System Administrator
Psychology Department, Rutgers University, Newark campus
Reply to: