Bug#732939: Re: Bug#732939: linux-image-3.11-2-686-pae: Data corruption with ASMedia ASM1061 SATA IDE Controller in AHCI-Mode
Sorry i just looked this up. I already replied to Ben but not to the bug. For now i can say that after a few weeks (5-10) one disk failed in the array. Maybe this caused the curruption? SMART did not warn me anyway. Heres what i wrote to Ben: Hey Ben, Sorry for the late answer. Did the kernel log any I/O error messages (not from ext4) while writing through the ASRock controller in AHCI mode? No there was nothing that warned me about the problem Do you have a single hard drive connected to each port or are you using a port multiplier? No multiplier in use. Now some thoughts about what happend: I was not able to recover the data. Fsck just went out of performance. It fixed 235M inodes of 450M and then i was only able to fix 1M inodes in 24h which was not worth the time so i stopped recovering and created a new filesystem. The problem seams to be gone and i cannot say why because i did various things. Its most likely that a bios update fixed it which would make this problem firmware related and not kernel related. Please close this bug its seams theres no problem with kernel. Maybe its fixed by updating the bios. I am not sure. Asmedia is running with AHCI enabled now and no more corruption. Kind regards, Daniel -- To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#732939: linux-image-3.11-2-686-pae: Data corruption with ASMedia ASM1061 SATA IDE Controller in AHCI-Mode
Hi Daniel, Ben Hutchings wrote (24 Dec 2013 02:54:26 GMT) : Did the kernel log any I/O error messages (not from ext4) while writing through the ASRock controller in AHCI mode? Do you have a single hard drive connected to each port or are you using a port multiplier? Do you think you'll have time to answer this request for additional information sent by Ben a bit more than two months ago? Also, it might be useful to try and reproduce this with Linux 3.13.x from Debian unstable, if possible. Cheers, -- intrigeri | GnuPG key @ https://gaffer.ptitcanardnoir.org/intrigeri/intrigeri.asc | OTR fingerprint @ https://gaffer.ptitcanardnoir.org/intrigeri/otr.asc -- To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Processed: Re: Bug#732939: linux-image-3.11-2-686-pae: Data corruption with ASMedia ASM1061 SATA IDE Controller in AHCI-Mode
Processing control commands: tag -1 moreinfo Bug #732939 [src:linux] linux-image-3.11-2-686-pae: Data corruption with ASMedia ASM1061 SATA IDE Controller in AHCI-Mode Added tag(s) moreinfo. -- 732939: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=732939 Debian Bug Tracking System Contact ow...@bugs.debian.org with problems -- To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#732939: linux-image-3.11-2-686-pae: Data corruption with ASMedia ASM1061 SATA IDE Controller in AHCI-Mode
Control: tag -1 moreinfo On Sun, 2013-12-22 at 22:22 +0100, Daniel Koch wrote: Package: src:linux Version: 3.11.10-1 Severity: critical Justification: causes serious data loss Dear Maintainer, As mentioned here http://www.linuxquestions.org/questions/linux-hardware-18/hard-drive-corruption-with-asm1062-sata-controller-4175469744/ by somebody else using the ASMedia in AHCI-Mode causes serious data corruption. Almost every mainboard should be configured using AHCI by default which makes this bug critical. I am currently facing this bug on my own. I replaced the mainboard in my NAS with an ASRock Z77 Pro4-M which has two SATA Ports using the asm1060 controller and was configured to use AHCI by the manufacture. After replacing the device i booted the system and everything looked fine. Some hours later i noticed the data corruption on the ext4-filesystem and i had to run fscks which took very long. The linux-raid device stores 12 TB and the system has 4 GB RAM which caused fsck to stop with cannot allocate memory so i had to started it again and again. I noticed that fsck was fixing the same inodes several timesy. After finding the threat on linuxquestions.org i set ACHI to IDE and fsck is now going forward and stopped to fix the same inodes again and again which makes me belive that the corruption has stopped. Also i wrote files with the same content and created sha1sums of them which do match . I dont want to check if the would differ when using AHCI. [...] Did the kernel log any I/O error messages (not from ext4) while writing through the ASRock controller in AHCI mode? Do you have a single hard drive connected to each port or are you using a port multiplier? Ben. -- Ben Hutchings Unix is many things to many people, but it's never been everything to anybody. signature.asc Description: This is a digitally signed message part
Bug#732939: linux-image-3.11-2-686-pae: Data corruption with ASMedia ASM1061 SATA IDE Controller in AHCI-Mode
Package: src:linux Version: 3.11.10-1 Severity: critical Justification: causes serious data loss Dear Maintainer, As mentioned here http://www.linuxquestions.org/questions/linux-hardware-18/hard-drive-corruption-with-asm1062-sata-controller-4175469744/ by somebody else using the ASMedia in AHCI-Mode causes serious data corruption. Almost every mainboard should be configured using AHCI by default which makes this bug critical. I am currently facing this bug on my own. I replaced the mainboard in my NAS with an ASRock Z77 Pro4-M which has two SATA Ports using the asm1060 controller and was configured to use AHCI by the manufacture. After replacing the device i booted the system and everything looked fine. Some hours later i noticed the data corruption on the ext4-filesystem and i had to run fscks which took very long. The linux-raid device stores 12 TB and the system has 4 GB RAM which caused fsck to stop with cannot allocate memory so i had to started it again and again. I noticed that fsck was fixing the same inodes several timesy. After finding the threat on linuxquestions.org i set ACHI to IDE and fsck is now going forward and stopped to fix the same inodes again and again which makes me belive that the corruption has stopped. Also i wrote files with the same content and created sha1sums of them which do match . I dont want to check if the would differ when using AHCI. Here are some actions fsck is doing right now: Inode 9943178, i_size is 11096869482525051813, should be 0. Fix? yes Inode 9943178, i_blocks is 196842723066601, should be 0. Fix? yes Inode 9943179 is in use, but has dtime set. Fix? yes Inode 9943179 has imagic flag set. Clear? yes Inode 9943179 has a extra size (26461) which is invalid Fix? yes Inode 9943179 has a bad extended attribute block 591467076. Clear? yes Inode 9943179 has INDEX_FL flag set but is not a directory. Clear HTree index? yes Inode 9943179, i_size is 10152020167559741440, should be 0. Fix? yes Inode 9943179, i_blocks is 171648422719524, should be 0. Fix? yes Inode 9943180 is in use, but has dtime set. Fix? yes Inode 9943180 has imagic flag set. Clear? yes Inode 9943180 has a extra size (44153) which is invalid Fix? yes Inode 9943180 has a bad extended attribute block 1308992673. Clear? yes Inode 9943180, i_size is 2358728759865762938, should be 0. Fix? yes Inode 9943180, i_blocks is 119552820074414, should be 0. Fix? yes Inode 9943181 is in use, but has dtime set. Fix? yes Inode 9943181 has imagic flag set. Clear? yes Inode 9943181 has a extra size (40687) which is invalid -- Package-specific info: ** Version: Linux version 3.11-2-686-pae (debian-ker...@lists.debian.org) (gcc version 4.8.2 (Debian 4.8.2-7) ) #1 SMP Debian 3.11.10-1 (2013-12-04) ** Command line: BOOT_IMAGE=/boot/vmlinuz-3.11-2-686-pae root=UUID=7a0defcb-ccf7-4bb4-91f1-74367d60c8be ro quiet pcie_aspm=force i915.i915_enable_fbc i915.lvds_downclock=1 ** Not tainted ** Kernel log: [17355.225214] EXT4-fs error (device md0): ext4_iget:4242: inode #81202494: comm smbd: bogus i_mode (151012) [17355.227106] EXT4-fs error (device md0): ext4_iget:4242: inode #81202387: comm smbd: bogus i_mode (57634) [17355.229019] EXT4-fs error (device md0): ext4_iget:4242: inode #81202560: comm smbd: bogus i_mode (162442) [17355.231272] EXT4-fs error (device md0): ext4_iget:4242: inode #81202493: comm smbd: bogus i_mode (166744) [17355.233222] EXT4-fs error (device md0): ext4_lookup:1437: inode #81215504: comm smbd: deleted inode referenced: 81202458 [17355.235122] EXT4-fs error (device md0): ext4_iget:4242: inode #81202610: comm smbd: bogus i_mode (130313) [17355.237065] EXT4-fs error (device md0): ext4_iget:4193: inode #81202415: comm smbd: bad extended attribute block 4049073417 [17355.239033] EXT4-fs error (device md0): ext4_iget:4242: inode #81202592: comm smbd: bogus i_mode (164142) [17355.241393] EXT4-fs error (device md0): ext4_lookup:1437: inode #81215504: comm smbd: deleted inode referenced: 81202471 [17355.243340] EXT4-fs error (device md0): ext4_iget:4242: inode #81202672: comm smbd: bogus i_mode (173123) [17355.245372] EXT4-fs error (device md0): ext4_iget:4242: inode #81202462: comm smbd: bogus i_mode (52117) [17355.247307] EXT4-fs error (device md0): ext4_lookup:1437: inode #81215504: comm smbd: deleted inode referenced: 81202406 [17355.249313] EXT4-fs error (device md0): ext4_iget:4193: inode #81202362: comm smbd: bad extended attribute block 4294967119 [17355.251482] EXT4-fs error (device md0): ext4_lookup:1437: inode #81215504: comm smbd: deleted inode referenced: 81202521 [17355.253523] EXT4-fs error (device md0): ext4_iget:4242: inode #81202618: comm smbd: bogus i_mode (132201) [17355.255537] EXT4-fs error (device md0): ext4_iget:4193: inode #81202586: comm smbd: bad extended attribute block 3518414528 [17355.257609] EXT4-fs error (device md0): ext4_lookup:1437: inode #81215504: comm smbd: deleted inode referenced: 81202379