Re: 5.2-CURRENT crashes everyday

On Tue, 10 Aug 2004, Robert Watson wrote:

> On Tue, 10 Aug 2004, Scott Long wrote:
>
> > I'm pretty sure that this is a known, harmless bug.  It might even have
> > been fixed in the last few weeks, but I can't remember for sure.  You
> > can avoid the panics by removing INVARIANTS from your kernel config.
>
> Lock reversals don't generate a panic by default, so the hang must happen
> shortly after the lock order reversal.  I.e., there's a real bug here, but
> it may or may not be related to the WITNESS output.
>
> Andrew -- if you use a serial console, are you able to use a serial break
> to get into the debugger if the debugger is compiled in with serial break
> support?  You can find details on setting up the debugging environment in
> the handbook.

apart from that I remember someone with exact the same symptoms and
the same report (no serial log but only this well known LOR) but I
cannot remember what had been the actual problem or if it got solved.
Search for this LOR in the archives of the last two months. Might help
anyone who is going to look into this to get more input.


> > Andrew A. Leikand wrote:
> > >
> > >  well i have a real big problem it is the server with 5.2-current.
> > > Harware is IBM eServer 345 Dual Xeon with serveRAID 6i, raid controller
> > > has no support under STABLE thus i had no choice :(
> > >  There are apache and sendmail on the server, load averages about 0.01,
> > > but it crashes everyday and i have no idea how to force it work.
> > >  The only messages before it goes down is
> > >
> > > lock order reversal
> > >  1st 0xc6574738 vm object (vm object) _at_ /usr/src/sys/vm/swap_pager.c:1311
> > >  2nd 0xc0673ae0 swap_pager swhash (swap_pager swhash) _at_ /usr/src/sys/vm/swap_pager.c:1797
> > >  3rd 0xc0c43a50 vm object (vm object) _at_ /usr/src/sys/vm/uma_core.c:925
> > > Stack backtrace:
> > > backtrace(0,1,c064cf90,c064e0c0,c06194dc) at backtrace+0x12
> > > witness_checkorder(c0c43a50,9,c05fe7bf,39d) at witness_checkorder+0x53b
> > > _mtx_lock_flags(c0c43a50,0,c05fe7bf,39d,c350b288) at _mtx_lock_flags+0x57
> > > obj_alloc(c3502dc0,1000,de086a2b,101,de086a38) at obj_alloc+0x31
> > > slab_zalloc(c3502dc0,1,c3502dc0,c3502dc0,c350b280) at slab_zalloc+0x87
> > > uma_zone_slab(c3502dc0,1,c350b288,0,c05fe7bf,79c) at uma_zone_slab+0xb0
> > > uma_zalloc_internal(c3502dc0,0,1,c350b288,0) at uma_zalloc_internal+0x29
> > > uma_zalloc_arg(c3502dc0,0,1) at uma_zalloc_arg+0x2a2
> > > swp_pager_meta_build(c6574738,5e,0,2,0) at swp_pager_meta_build+0x108
> > > swap_pager_putpages(c6574738,de086bf0,1,0,de086b60) at swap_pager_putpages+0x2a8
> > > default_pager_putpages(c6574738,de086bf0,1,0,de086b60) at default_pager_putpages+0x18
> > > vm_pageout_flush(de086bf0,1,0,c064c6e0,2ff) at vm_pageout_flush+0x112
> > > vm_pageout_clean(c2ca1f88) at vm_pageout_clean+0x2a5
> > > vm_pageout_scan(0,c0673fe0,0,c05fe55f,5a7) at vm_pageout_scan+0x543
> > > vm_pageout(0,de086d48,0,c057f0e4,0) at vm_pageout+0x2cf
> > > fork_exit(c057f0e4,0,de086d48) at fork_exit+0x98
> > > fork_trampoline() at fork_trampoline+0x8
> > > --- trap 0x1, eip = 0, esp = 0xde086d7c, ebp = 0 ---
> > >
> > >  Kernel config and dmesg  are attached.
> > >
> > > Appreciate any comments or feedback on this.

-- 
Bjoern A. Zeeb				bzeeb at Zabbadoz dot NeT