Ext4 kernel panic

Question

I'm getting such kernel panics:

    EXT4-fs error (device md2): ext4_ext_find_extent: bad header/extent in inode #97911179: invalid magic - magic 5f69, entries 28769, max 26988(0), depth 24939(0)
----------- [cut here ] --------- [please bite here ] ---------
Kernel BUG at fs/ext4/extents.c:1973
invalid opcode: 0000 [1] SMP
last sysfs file: /devices/pci0000:00/0000:00:00.0/irq
CPU 6
Modules linked in: iptable_filter ipt_REDIRECT ip_nat_ftp ip_conntrack_ftp iptable_nat ip_nat ip_tables xt_state ip_conntrack_netbios_ns ip_conntrack nfnetlink netconsole ipt_iprange xt_tcpudp autofs4 hwmon_vid coretemp cpufreq_ondemand acpi_cpufreq freq_table mperf x_tables be2iscsi ib_iser rdma_cm ib_cm iw_cm ib_sa ib_mad ib_core ib_addr iscsi_tcp bnx2i cnic ipv6 xfrm_nalgo crypto_api uio cxgb3i libcxgbi cxgb3 8021q libiscsi_tcp libiscsi2 scsi_transport_iscsi2 scsi_transport_iscsi ext3 jbd dm_mirror dm_multipath scsi_dh video backlight sbs power_meter hwmon i2c_ec dell_wmi wmi button battery asus_acpi acpi_memhotplug ac lp joydev sg shpchp parport_pc parport r8169 mii serio_raw tpm_tis tpm tpm_bios i2c_i801 i2c_core pcspkr dm_raid45 dm_message dm_region_hash dm_log dm_mod dm_mem_cache raid10 raid456 xor raid0 sata_nv aacraid 3w_9xxx 3w_xxxx sata_sil sata_via ahci libata sd_mod scsi_mod raid1 ext4 jbd2 crc16 uhci_hcd ohci_hcd ehci_hcd
Pid: 9374, comm: httpd Not tainted 2.6.18-308.20.1.el5debug 0000001
RIP: 0010:[<ffffffff8806ccda>] [<ffffffff8806ccda>] :ext4:ext4_ext_put_in_cache+0x21/0x6a
RSP: 0018:ffff8101c2df7678 EFLAGS: 00010246
RAX: 00000000fffffbf1 RBX: ffff810758115dc8 RCX: 0000000000000000
RDX: 0000000000000000 RSI: 0000000000000000 RDI: ffff810758115958
RBP: ffff810758115958 R08: 0000000000000002 R09: 0000000000000000
R10: ffff8101c2df75a0 R11: 0000000000000100 R12: 0000000000000000
R13: 0000000000000002 R14: 0000000000000000 R15: 0000000000000000
FS: 00002ab948d31f70(0000) GS:ffff81081f4ba4c8(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
CR2: 000000001de9e4e0 CR3: 000000014ae88000 CR4: 00000000000006a0
Process httpd (pid: 9374, threadinfo ffff8101c2df6000, task ffff8101cdf74d80)
Stack: 000181070000040f ffff810758115dc8 ffff8103f15d7ff4 ffff8107581157f0
 ffff810758115958 000000000000040f 0000000000000000 ffffffff8806f621
 ffff8101c2df76d8 ffff8101c2df7738 0000000000000000 ffff81034900c310
Call Trace:
 [<ffffffff8806f621>] :ext4:ext4_ext_get_blocks+0x258/0x16f3
 [<ffffffff80013994>] poison_obj+0x26/0x2f
 [<ffffffff800331e2>] cache_free_debugcheck+0x20b/0x21a
 [<ffffffff8805b4ac>] :ext4:ext4_get_blocks+0x43/0x1d2
 [<ffffffff8805b4cf>] :ext4:ext4_get_blocks+0x66/0x1d2
 [<ffffffff8805c16a>] :ext4:ext4_get_block+0xa7/0xe6
 [<ffffffff8805c3be>] :ext4:ext4_block_truncate_page+0x215/0x4f1
 [<ffffffff8806e832>] :ext4:ext4_ext_truncate+0x65/0x909
 [<ffffffff8805b4f9>] :ext4:ext4_get_blocks+0x90/0x1d2
 [<ffffffff8805ccfc>] :ext4:ext4_truncate+0x91/0x53b
 [<ffffffff80041e5d>] pagevec_lookup+0x17/0x1e
 [<ffffffff8002d3cf>] truncate_inode_pages_range+0x1f3/0x2d5
 [<ffffffff8803b78b>] :jbd2:jbd2_journal_stop+0x1f1/0x201
 [<ffffffff8805f3c1>] :ext4:ext4_da_write_begin+0x1ea/0x25b
 [<ffffffff80010896>] generic_file_buffered_write+0x151/0x6c3
 [<ffffffff800174b1>] __generic_file_aio_write_nolock+0x36c/0x3b9
 [<ffffffff800482ab>] do_sock_read+0xcf/0x110
 [<ffffffff80022d49>] generic_file_aio_write+0x69/0xc5
 [<ffffffff88056c0a>] :ext4:ext4_file_write+0xcb/0x215
 [<ffffffff8001936b>] do_sync_write+0xc7/0x104
 [<ffffffff8000d418>] dnotify_parent+0x1f/0x7b
 [<ffffffff800efead>] do_readv_writev+0x26e/0x291
 [<ffffffff800a8192>] autoremove_wake_function+0x0/0x2e
 [<ffffffff80035b9f>] do_setitimer+0x62a/0x692
 [<ffffffff8002e6a5>] mntput_no_expire+0x19/0x8d
 [<ffffffff80049aa0>] sys_chdir+0x55/0x62
 [<ffffffff800178c6>] vfs_write+0xce/0x174
 [<ffffffff800181ba>] sys_write+0x45/0x6e
 [<ffffffff80060116>] system_call+0x7e/0x83


Code: 0f 0b 68 3e 27 08 88 c2 b5 07 eb fe 48 8d 9f 08 05 00 00 48
RIP [<ffffffff8806ccda>] :ext4:ext4_ext_put_in_cache+0x21/0x6a
 RSP <ffff8101c2df7678>
 <0>Kernel panic - not syncing: Fatal exception
 <0>Rebooting in 1 seconds..

My system is CentOS 5.8 64-bit.

/dev/md2        /home   ext4    rw,noatime,nodiratime,usrjquota=aquota.user,grpjquota=aquota.group,usrquota,grpquota,jqfmt=vfsv0     0       0

Kernel: 2.6.18-308.20.1.el5debug

md2 : active raid1 sdc3[0] sdd3[1]
      2914280100 blocks super 1.0 [2/2] [UU]
      [>....................]  resync =  0.2% (7252288/2914280100) finish=13468.3min speed=3595K/sec

/dev/md2              2,7T  1,8T  908G  67% /home

How can I fix this ?

Even when resync my array, unmount filesystems and then check and fix all errors it all works fine for about a week of time and then small ext4 errors showing and finally it starts kernel panicing.

Answer 1

1. Wait for the RAID resync to complete. Since it's started, if it can run to completion, that'll be best.
2. Reboot the system in single-user mode, unmount the /home partition (if it's mounted) and do an e2fsck -f /dev/md2, to ensure the file system is self-consistent.
3. If, after that, this happens again, it's likely indicative either of bad hardware or of a genuine kernel bug. If you can't find any evidence of the former, make sure you're running the very latest C5 kernel with a yum update, and if so, log a bug with the CentOS project via their bug tracker (after searching to ensure this is not a known issue).

Edit: I wouldn't regress the kernel without compelling evidence that there's some known issue with the current kernel. If your kernel is logging continuous, progressive FS corruption then for me that is very strongly indicative of hardware issues.

Have you run smartctl checks on the sdc and sdd discs? You say "the discs are fine", but you don't say how you know that.

If the discs really are fine, then I notice that you're using only a partition on sdc and sdd to provide the metadevice - something that's worth checking is that the partition tables don't overlap. I've known problems be caused when partitions overlapped by a few blocks, because the superblock at the bottom of one file system kept treading on the blocks at the very top of the other FS.

Edit 2: thanks for the smartctl output. Unfortunately, the "health check is passed" output is fairly meaningless, because neither disc has ever been tested ("No self-tests have been logged"). Try a smartctl -t long /dev/sdc, and when it's finished, the same for sdd, and then see what smartctl says.

Answer 2

You are a victim of file system corruption. However, you have a nasty configuration in your /etc/fstab which prevents any type of check of that file system. Please read man fstab, specifically this section about the sixth field:

   The sixth field (fs_passno).
          This  field  is used by the fsck(8) program to determine the order
          in which filesystem checks are done  at  reboot  time.   The  root
          filesystem  should  be  specified with a fs_passno of 1, and other
          filesystems should have a fs_passno of 2.   Filesystems  within  a
          drive  will  be checked sequentially, but filesystems on different
          drives will be checked at the same  time  to  utilize  parallelism
          available  in  the hardware.  If the sixth field is not present or
          zero, a value of zero is returned and fsck will  assume  that  the
          filesystem does not need to be checked.

So, you've told the system not to fsck this file system ever. Please just check and fix your filesystem offline (fsck). It has corrupted.

Answer 3

I've installed this kernel: http://elrepo.org/linux/kernel/el5/x86_64/RPMS/ then fixed filesystem and resynced arrays and problem is gone.