Hi, I have set up a btrfs within a 200 GB file that I mount via -o loop. It worked fine so far but today it crashed when there was alot of concurrent writing/reading going on. I''m testing it with a torrent client, so the data is really irrelevant - but it''s a good amount of stress. When it crashed dmesg got logged onto my /var/log which filled up the entire harddrive within a few minutes... Since I wasn''t able to reboot properly I deleted the relevant log files and am not able to give a dmesg output of the crash :( I adjusted my logging now so I should be able to provide more info next time. When I try to mount the filesystem now mount hangs forever though and dmesg is flooding this message several times per second: parent transid verify failed on 164143104 wanted 22325 found 22209 I don''t really care about the data but I would like to restore one config file from the filesystem. Is there any way to get it mounted? Or to extract single files? If you need any more info feel free to ask, I''ll try to help as much as possible. Thanks Harald -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Sun, 2009-04-12 at 02:03 +0200, Harald Glatt wrote:> Hi, > > I have set up a btrfs within a 200 GB file that I mount via -o loop. > > It worked fine so far but today it crashed when there was alot of > concurrent writing/reading going on. > I''m testing it with a torrent client, so the data is really irrelevant > - but it''s a good amount of stress. >Loopback is somewhat tricky because the kernel loopback driver only writes into the page cache of the underlying file in the host filesystem. So, the filesystem inside the loopback file thinks it has written things to disk but in reality it is just in cache. If you crash in the middle of all of this, you''re going to see significant corruptions of the FS inside the loopback file. -chris -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Mon, Apr 13, 2009 at 09:28:28AM -0400, Chris Mason wrote:> On Sun, 2009-04-12 at 02:03 +0200, Harald Glatt wrote: > > Hi, > > > > I have set up a btrfs within a 200 GB file that I mount via -o loop. > > > > It worked fine so far but today it crashed when there was alot of > > concurrent writing/reading going on. > > I''m testing it with a torrent client, so the data is really irrelevant > > - but it''s a good amount of stress. > > > > Loopback is somewhat tricky because the kernel loopback driver only > writes into the page cache of the underlying file in the host > filesystem. So, the filesystem inside the loopback file thinks it has > written things to disk but in reality it is just in cache. > > If you crash in the middle of all of this, you''re going to see > significant corruptions of the FS inside the loopback file.There''s a patch floating around to add barrier support to the loop driver to fix this issue. I hoped it would get merged for 2.6.30 but it''s not in yet. -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html