https://bugzilla.novell.com/show_bug.cgi?id=471382 User robin.listas@telefonica.net added comment https://bugzilla.novell.com/show_bug.cgi?id=471382#c1 Carlos Robinson changed: What |Removed |Added ---------------------------------------------------------------------------- Summary|External USB disk |External USB disk |(encrypted, reiserfs) gets |(reiserfs) gets corrupted |corrupted and becomes |and becomes read-only. |read-only. | --- Comment #1 from Carlos Robinson 2009-02-01 20:36:28 MST --- Some time after I wrote the above, partition /dev/sda1 and sda2 also became bad (without disk activity that I know of): Feb 2 02:10:49 nimrodel kernel: sd 0:0:0:0: [sda] Sense Key : No Sense [current] Feb 2 02:10:49 nimrodel kernel: Info fld=0x0 Feb 2 02:10:49 nimrodel kernel: sd 0:0:0:0: [sda] Add. Sense: No additional sense information Feb 2 02:10:49 nimrodel kernel: REISERFS warning: reiserfs-5090 is_tree_node: node level 29550 does not match to the expected one 1 Feb 2 02:10:49 nimrodel kernel: REISERFS error (device sda1): vs-5150 search_by_key: invalid format found in block 14286849. Fsck? Feb 2 02:10:49 nimrodel kernel: REISERFS (device sda1): Remounting filesystem read-only Feb 2 02:11:45 nimrodel kernel: sd 0:0:0:0: [sda] Sense Key : No Sense [current] Feb 2 02:11:45 nimrodel kernel: Info fld=0x0 Feb 2 02:11:45 nimrodel kernel: sd 0:0:0:0: [sda] Add. Sense: No additional sense information Feb 2 02:11:45 nimrodel kernel: REISERFS warning: reiserfs-5090 is_tree_node: node level 60651 does not match to the expected one 1 Feb 2 02:11:45 nimrodel kernel: REISERFS error (device sda2): vs-5150 search_by_key: invalid format found in block 12648508. Fsck? Feb 2 02:11:45 nimrodel kernel: REISERFS (device sda2): Remounting filesystem read-only and a lot more entries. Thus this bug is not related to encryption as thought before. Changing tittle. Notice that I have used this same disk on 10.3 and was not problematic: problems started just after the upgrade to 11.0. It is my current external backup disk, bought previous to the 11.0 upgrade, and is not much used. Running reiserfck now (4:20 AM CET) -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=471382 User robin.listas@telefonica.net added comment https://bugzilla.novell.com/show_bug.cgi?id=471382#c3 --- Comment #3 from Carlos Robinson 2009-02-02 19:55:47 MST --- Created an attachment (id=269466) --> (https://bugzilla.novell.com/attachment.cgi?id=269466) log and notes of the recovery procedure These are the operations I had to do to recover the 4 partitions. It took over 18 hours, sleep period included. I have seen a post in the mailing list for a similar problem (I think), but under vmware: they say the system has not time to write the journal to disk because it is too slow (the host disk is busy), and it becomes read-only on guest. Related to scsi disk system (subject: Volume mount going mad.. Date: Mon, 2 Feb 2009 15:00:48 +0100) -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=471382 User dominique.leuenberger@tmf-group.com added comment https://bugzilla.novell.com/show_bug.cgi?id=471382#c4 Dominique Leuenberger changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |dominique.leuenberger@tmf-g | |roup.com --- Comment #4 from Dominique Leuenberger 2009-02-20 06:43:33 MST --- I see similiar things on my system: for my system the root partition goes to read-only. FSck needed (and, as it's the root partition, also a reboot). Did not see similiar things happening on 10.2. Difference between my situation and Carlos': - I'm running openSUSE 11.1 - My partitions are EXT3 (on LVM) Not sure how relevant this report is, but they might have similiar sources. (I changed already the disk and it did not disappear...) -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=471382 User robin.listas@telefonica.net added comment https://bugzilla.novell.com/show_bug.cgi?id=471382#c6 Carlos Robinson changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |NEW Info Provider|robin.listas@telefonica.net | --- Comment #6 from Carlos Robinson 2009-02-22 05:50:05 MST --- Well... This problem can affect a disk and not be noticed till the next time I use it; its difficult to know if it has appeared with the current session. In fact, I used it recently in 11.0 and there hasn't been any entry in the kernel log. I don't know if it has been chance, or because there was a kernel update (Jan 28). I had the very wild idea of changing my connect procedure. Normally I leave the usb cable connected, and then I power up the disk. Now I'm connecting the usb cable after powering up the disk; well, the weird thing is that the two times I have used the disk with this procedure I have seen no suspicious log entries. Yesterday I booted my 11.1 test partition, which has kernel kernel-debug-base-2.6.27.13-SL111_BRANCH_20090203134609_5784a3e1, and I copied over to the usb disk about 15..20 GiB of big files, and I saw no suspicious entries in the log. I even had done: echo "set enabled=1 usbcore" > /sys/kernel/debug/dynamic_printk/modules just in case (from a recommendation in 343874#c6). Later I rebooted to 11.0 and mounted the disk, with no errors either. And of course, this was a "brief" test, I can't update to 11.1 and see what happens after weeks of usage. So... I don't know. The connection procedure? New kernel? That the problematic operation is not write file but rename files (which I did not do recently)? I'm baffled. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=471382 User robin.listas@telefonica.net added comment https://bugzilla.novell.com/show_bug.cgi?id=471382#c8 --- Comment #8 from Carlos Robinson 2009-02-25 19:46:33 MST --- Big corruption. reiserfsck dumps hundreds of lines and ends by requesting --rebuild-tree. .. bad_indirect_item: block 403894: The item [321 489 0x1ac4001 IND (1)] has the bad pointer (1009) to the block (1725747481) bad_indirect_item: block 403894: The item [321 489 0x1ac4001 IND (1)] has the bad pointer (1010) to the block (2951855736) bad_indirect_item: block 403894: The item [321 489 0x1ac4001 IND (1)] has the bad pointer (1011) to the block (3041029050) / 81 (of 99|/ 10 (of 170-block 3213735: The level of the node (51419) is not correct, (1) expected the problem in the internal node occured (3213735), whole subtree is skipped / 82 (of 99\/ 89 (of 134|block 3557713: The level of the node (20720) is not correct, (1) expected the problem in the internal node occured (3557713), whole subtree is skipped / 90 (of 99\/ 35 (of 170-bad_indirect_item: block 17246368: The item (345 347 0x82a3001 IND (1), len 4048, location 48 entry count 0, fsck need 0, format new) has the bad pointer (297) to the block (17246659), which is in tree already / 96 (of 99//122 (of 170\bad_indirect_item: block 18368505: The item (345 376 0x12cde001 IND (1), len 4048, location 48 entry count 0, fsck need 0, format new) has the bad pointer (878) to the block (18369429), which is in tree already finished Comparing bitmaps..vpf-10640: The on-disk and the correct bitmaps differs. Bad nodes were found, Semantic pass skipped 4 found corruptions can be fixed only when running with --rebuild-tree ########### reiserfsck finished at Thu Feb 26 01:58:28 2009 ########### -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=471382 User robin.listas@telefonica.net added comment https://bugzilla.novell.com/show_bug.cgi?id=471382#c10 --- Comment #10 from Carlos Robinson 2009-03-05 18:04:54 MST --- More info. Last time I said I was going to mount the device in r/w mode only on 11.1, and r-only on 11.0. The problem is that, in ro mode I have errors, which would mean that the corruption was written from 11.1 :-( I was copying 4 GB of data in big files. While reading, one of them failed: Mar 6 00:11:58 nimrodel kernel: sd 1:0:0:0: [sda] Attached SCSI disk Mar 6 00:11:58 nimrodel kernel: sd 1:0:0:0: Attached scsi generic sg0 type 0 Mar 6 00:11:58 nimrodel kernel: usb-storage: device scan complete Mar 6 00:12:22 nimrodel kernel: REISERFS (device dm-5): found reiserfs format "3.6" with standard journal Mar 6 00:12:22 nimrodel kernel: REISERFS (device dm-5): using ordered data mode Mar 6 00:12:22 nimrodel kernel: REISERFS (device dm-5): journal params: device dm-5, size 8192, journal first block 18, max trans len 1024, max batch 900, max commit age 30, max trans age 30 Mar 6 00:12:22 nimrodel kernel: REISERFS (device dm-5): checking transaction log (dm-5) Mar 6 00:12:22 nimrodel kernel: REISERFS (device dm-5): Using r5 hash to sort names Mar 6 00:19:04 nimrodel kernel: sd 1:0:0:0: [sda] Sense Key : No Sense [current] Mar 6 00:19:04 nimrodel kernel: Info fld=0x0 Mar 6 00:19:04 nimrodel kernel: sd 1:0:0:0: [sda] Add. Sense: No additional sense information Mar 6 00:19:04 nimrodel kernel: REISERFS warning: reiserfs-5090 is_tree_node: node level 5613 does not match to the expected one 1 Mar 6 00:19:04 nimrodel kernel: REISERFS error (device dm-5): vs-5150 search_by_key: invalid format found in block 13374638. Fsck? Mar 6 00:19:04 nimrodel kernel: REISERFS warning: reiserfs-5090 is_tree_node: node level 5613 does not match to the expected one 1 Mar 6 00:19:04 nimrodel kernel: REISERFS error (device dm-5): vs-5150 search_by_key: invalid format found in block 13374638. Fsck? Mar 6 00:19:04 nimrodel kernel: REISERFS warning: reiserfs-5090 is_tree_node: node level 5613 does not match to the expected one 1 Mar 6 00:19:04 nimrodel kernel: REISERFS error (device dm-5): vs-5150 search_by_key: invalid format found in block 13374638. Fsck? Mar 6 00:19:04 nimrodel kernel: REISERFS warning: reiserfs-5090 is_tree_node: node level 5613 does not match to the expected one 1 Mar 6 00:19:04 nimrodel kernel: REISERFS error (device dm-5): vs-5150 search_by_key: invalid format found in block 13374638. Fsck? Mar 6 00:19:04 nimrodel kernel: REISERFS warning: reiserfs-5090 is_tree_node: node level 5613 does not match to the expected one 1 Mar 6 00:19:04 nimrodel kernel: REISERFS error (device dm-5): vs-5150 search_by_key: invalid format found in block 13374638. Fsck? and many more like that. I could not copy that file. Then I umounted and mounted the device, and copied the file sucesfully: Mar 6 00:41:09 nimrodel kernel: REISERFS (device dm-5): found reiserfs format "3.6" with standard journal Mar 6 00:41:09 nimrodel kernel: REISERFS (device dm-5): using ordered data mode Mar 6 00:41:09 nimrodel kernel: REISERFS (device dm-5): journal params: device dm-5, size 8192, journal first block 18, max trans len 1024, max batch 900, max commit age 30, max trans age 30 Mar 6 00:41:09 nimrodel kernel: REISERFS (device dm-5): checking transaction log (dm-5) Mar 6 00:41:09 nimrodel kernel: REISERFS (device dm-5): Using r5 hash to sort names So... could the bug provoke a _read_ error, while the data is correct on the disk? -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.