[Mauiusers] maui crash

Dave Jackson jacksond at clusterresources.com
Fri Oct 14 20:49:59 MDT 2005


Andrew,

  A new Maui snapshot has been released.  This snapshot contains a fix
to prevent the crash you saw and also extends Maui allocation management
support for reservation based allocation charges with GOLD.  

  Please let us know if you see any further issues.

Thanks,
Dave

On Tue, 2005-10-11 at 14:56 -0400, Andrew J Caird wrote:
> That explains why I didn't see it repeatedly; I was testing reservation 
> based charging at the time.  I gave up when it didn't seem to work, but 
> I'll try again when I get the latest snapshot installed.
> 
> Is there an announcement list for snapshot releases?  I don't seem to hear 
> when they come out.  Also, is there a list of changes from snapshot to 
> snapshot?
> 
> Thanks.
> --andy
> 
> On Tue, 11 Oct 2005, Dave Jackson wrote:
> 
> > Andrew,
> >
> >  Maui had no support for reservation based charging with Gold.  The
> > failure you saw should only occur when a reservation completes which has
> > one or more accountable credentials associated with it.  The new changes
> > are complete and are going through regression testing now.  Thanks for
> > finding it and please let us know if you see anything further.
> >
> > Thanks,
> > Dave
> >
> > On Tue, 2005-10-11 at 13:12 -0400, Andrew J Caird wrote:
> >> Dave,
> >>
> >> Thanks.  I hope you aren't rushing for me; like I said, I've only seen it
> >> a couple times (and was lucky, as it turns out, to get one of them in the
> >> debugger).
> >>
> >> Best,
> >> --andy
> >>
> >> On Tue, 11 Oct 2005, Dave Jackson wrote:
> >>
> >>> Andrew,
> >>>
> >>>  This looks like a pure bug.  We are fixing it now.  We hope to have a
> >>> snapshot available within the hour.
> >>>
> >>> Dave
> >>>
> >>> On Tue, 2005-10-11 at 10:06 -0400, Andrew J Caird wrote:
> >>>> Hello,
> >>>>
> >>>> I'm seeing a crash in maui (v3.2.6p14-snap.1126543721) with accounting
> >>>> enabled.  I haven't tried to debug it too much, but the crash looks
> >>>> like:
> >>>>
> >>>> (gdb) run -d
> >>>>          Keeping /usr/local/maui-3.2.6p14-snap/sbin/maui...
> >>>> Starting program: /usr/local/maui-3.2.6p14-snap/sbin/maui -d
> >>>>
> >>>> Program received signal SIGSEGV, Segmentation fault.
> >>>> 0x00000000400f11f4 in MXMLDestroyE (EP=0x7fbfff2dc8) at MXML.c:320
> >>>> 320           if (E->C[index] == NULL)
> >>>> (gdb) where
> >>>> #0  0x00000000400f11f4 in MXMLDestroyE (EP=0x7fbfff2dc8) at MXML.c:320
> >>>> #1  0x000000004003fe44 in MAMAllocRDebit (AM=0x40848440, R=0x42c51040,
> >>>> RIndex=0x7fbfff501c, ErrMsg=0x0)
> >>>>      at MAM.c:977
> >>>> #2  0x000000004005369d in MResChargeAllocation (RS=0x42c51040, Mode=2)
> >>>> at MRes.c:3820
> >>>> #3  0x00000000400564b4 in MResDestroy (RP=0x7fbfff5528) at MRes.c:5075
> >>>> #4  0x0000000040054e90 in MResCheckStatus (SR=0x0) at MRes.c:4535
> >>>> #5  0x00000000400035c0 in main (ArgC=2, ArgV=0x7fbffff648) at
> >>>> Server.c:204
> >>>> #6  0x0000002a957ea087 in __libc_start_main () from /lib64/libc.so.6
> >>>> #7  0x0000000040002d3a in _start ()
> >>>> (gdb) quit
> >>>>
> >>>> If I comment out my AMCFG line:
> >>>>       AMCFG[bank] TYPE=GOLD HOST=bank-host PORT=7112
> >>>>       CHARGEPOLICY=DEBITALLWC
> >>>>       JOBFAILUREACTION=NONE TIMEOUT=15 WIREPROTOCOL=XML SOCKETPROTOCOL=HTTP
> >>>>
> >>>> maui seems to run just fine; if I then stop maui and uncomment that
> >>>> line, maui continues to run just fine.  This has only happened twice,
> >>>> so it looks like some sort of job is causing a problem between maui and
> >>>> Gold.
> >
> >



More information about the mauiusers mailing list