Home » Mailing list archives » dragonflybsd-kernel » 2010 » April

dragonflybsd-kernel mailing list

FromSubjectsort iconDate
Matthew Dillon
Re: I/O errors on Hammer volume
At around the time when you noticed the problem, and/or a few days before or after, did you run a manual hammer cleanup and is it possible (if you did) that it might have been running at the same time as another hammer cleanup, such as the one run by cron? -Matt
[ message continues ]
Apr 16, 2:19 pm 2010
Matthew Dillon
Re: I/O errors on Hammer volume
I think I found the smoking gun but I can't be sure until I see the show output from Francois. I noticed that all the bad CRC'd records were typically around element 30-32 in the B-Tree node (out of ~64 elements). That is, the middle of the node. This implies a race between the reblocker/rebalancer and a node split during an insertion, or a race between the reblocker and the rebalancer. I am testing a fix now and I am not 100% sure that this was the issue, ...
[ message continues ]
Apr 16, 3:02 pm 2010
Matthew Dillon
Re: I/O errors on Hammer volume
I have pushed a new hammer directive to HEAD called 'checkmap'. It isn't complete yet but it should be good enough to verify that all record data references are accounted for in the blockmap. I still need to have it check B-Tree node allocations and other stuff. hammer -f <device> checkmap If it does not output any blockmap info then everything that it checks (so far) is ok. If you specify verbose (-v) it will output the blockmap info it collects ...
[ message continues ]
Apr 15, 8:52 pm 2010
previous daytodaynext day
April 15, 2010April 16, 2010April 17, 2010