Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 24 Mar 2004 09:37:21 -0500
From:      Don Bowman <don@sandvine.com>
To:        "'current@freebsd.org'" <current@freebsd.org>
Subject:   LOR on current
Message-ID:  <FE045D4D9F7AED4CBFF1B3B813C85337045D86E4@mail.sandvine.com>

next in thread | raw e-mail | index | archive | help

So my machine which has been locking up daily under current
(cvs up from saturday) this time showed this message to 
serial before locking up:

$ lock order reversal
 1st 0xc8f07084 vm object (vm object) @ /usr/src/sys/vm/swap_pager.c:1313
 2nd 0xc07738c0 swap_pager swhash (swap_pager swhash) @
/usr/src/sys/vm/swap_pager.c:1803
 3rd 0xc8beb294 vm object (vm object) @ /usr/src/sys/vm/uma_core.c:886
Stack backtrace:
backtrace(0,1,c07422e8,c0743490,c06feafc) at backtrace+0x12
witness_checkorder(c8beb294,9,c06d7841,376) at witness_checkorder+0x593
_mtx_lock_flags(c8beb294,0,c06d7841,376,c8bcb8d4) at _mtx_lock_flags+0x67
obj_alloc(c8bcb8c0,1000,ebefda3b,101,c8bcb8d4) at obj_alloc+0x31
slab_zalloc(c8bcb8c0,1,c8bcb8c0,c8bcb8d4,1) at slab_zalloc+0x82
uma_zone_slab(c8bcb8c0,1,c8bcb8d4,0,c06d7841,68f) at uma_zone_slab+0xb5
uma_zalloc_internal(c8bcb8c0,0,1,c8bcb8d4,0) at uma_zalloc_internal+0x28
uma_zalloc_arg(c8bcb8c0,0,1) at uma_zalloc_arg+0x2f8
swp_pager_meta_build(c8f07084,b,0,2,0) at swp_pager_meta_build+0x108
swap_pager_putpages(c8f07084,ebefdc00,1,0,ebefdb70) at
swap_pager_putpages+0x2b4
default_pager_putpages(c8f07084,ebefdc00,1,0,ebefdb70) at
default_pager_putpages+0x18
vm_pageout_flush(ebefdc00,1,0,0,ebefdbdc) at vm_pageout_flush+0x11e
vm_pageout_clean(c46590d8) at vm_pageout_clean+0x2a5
vm_pageout_scan(0) at vm_pageout_scan+0x54b
vm_pageout(0,ebefdd48,0,c0644404,0) at vm_pageout+0x2d2
fork_exit(c0644404,0,ebefdd48) at fork_exit+0xa8
fork_trampoline() at fork_trampoline+0x8

Its completely locked up now, so i can't get more information.
Does this tell anyone what they need to know to suggest where to make
a change?

The machine is running (only) postgresql. Doing two vacuum in parallel
will normally cause this.



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?FE045D4D9F7AED4CBFF1B3B813C85337045D86E4>