Bug #2358

DFBSD v3.0.2.32.g928ca - panic: hammer: insufficient undo FIFO space!

Added by tuxillo over 2 years ago. Updated about 2 years ago.

Status:In ProgressStart date:04/29/2012
Priority:NormalDue date:
Assignee:tuxillo% Done:

0%

Category:-
Target version:-

Description

Hi,

HAMMER filesystem recently created on a LVM stripe of 5 disks. Box has 1GB with no swap.

databank# lvs
LV VG Attr LSize Origin Snap% Move Log Copy% Convert
lv_root main -wi-a- 186.17g

I got no dump but I have located which panic in the code was hit: http://pkgbox64.dragonflybsd.org/xref/DragonFly-master/sys/vfs/hammer/hammer_redo.c#91

Backtrace function call is:

panic()
panic()
hammer_generate_redo()
hammer_vop_write()
vop_write()
vn_write()
kern_pwritev()
sys_extpwrite()
syscall2()
Xint0x80_syscall()

Image: http://leaf.dragonflybsd.org/~tuxillo/archive/pics/panic_fifo.jpg

After rebooting and trying to mount again the HAMMER FS:

databank# mount -a
hammer: mount on /mnt/safe: Result too large

And dmesg shows:

HAMMER(SAFE) recovery check seqno=00d791fe
HAMMER(SAFE) recovery range 30000000070ef490-3000000004a90000
HAMMER(SAFE) recovery nexto 3000000004a90000 endseqno=00e638f5
HAMMER(SAFE) recovery undo 30000000070ef490-3000000004a90000 (488246128 bytes)(RW)
HAMMER(SAFE) Found REDO_SYNC 30000000070ef490
HAMMER(SAFE) Ignoring extra REDO_SYNC records in UNDO/REDO FIFO.
HAMMER(SAFE) recovery complete
HAMMER(SAFE) recovery redo 30000000070ef490-3000000004a90000 (488246128 bytes)(RW)
HAMMER(SAFE) Find extended redo 30000000070ef490, 0 extbytes
HAMMER(SAFE) Find extended redo failed 34, unable to run REDO
HAMMER(SAFE) End redo recovery

I've activated swap and I'm going to try to reproduce it again and see if I can get a dump, but any tip would be appreciated :)

Cheers,
Antonio Huete


Related issues

Related to Bug #2402: Showstopper panics for Release 3.2 New 08/15/2012

History

#1 Updated by tuxillo about 2 years ago

  • Status changed from New to In Progress
  • Assignee set to tuxillo

Grab.

Also available in: Atom PDF