FedoraForum.org - Fedora Support Forums and Community
Results 1 to 7 of 7
  1. #1
    Join Date
    May 2009
    Posts
    15

    Fedora 14 - Software RAID issue

    Hey all!

    I've recently outgrown my 1TB Mirrored array, and have built an older box I had lying around as a file server. The box is configured as an nfs/samba server. I've bought two 2TB WD Green Drives, and a SATA controller to throw in for the mirror. I've opted to use Linux Software mirroring as I've had good luck with it in the past.

    However, I'm now facing a rather odd issue. Every time I reboot the box one of the drives fail to bind to the array. So it starts up with a single disk -- the second shows up as removed. The disk isn't always the same either. Sometimes it's /dev/sdb, sometimes it's /dev/sda. I can see the disk just fine. And when I do an mdadm --add /dev/md0 /dev/sd[a-b]1, it rebuilds sucessfully. I added a bitmap so the rebuild actually happens almost instantaneously now.

    Now -- this array is not bootable. I'm using a 36 GB disk that's comes in as /dev/sdc as the OS disk.

    /proc/mdstat:
    Personalities : [raid1]
    md0 : active raid1 sdb1[1]
    1953512400 blocks super 1.2 [2/1] [_U]
    bitmap: 2/15 pages [8KB], 65536KB chunk

    unused devices: <none>
    /dev/md0:
    Version : 1.2
    Creation Time : Fri Nov 19 08:42:53 2010
    Raid Level : raid1
    Array Size : 1953512400 (1863.01 GiB 2000.40 GB)
    Used Dev Size : 1953512400 (1863.01 GiB 2000.40 GB)
    Raid Devices : 2
    Total Devices : 1
    Persistence : Superblock is persistent

    Intent Bitmap : Internal

    Update Time : Tue Nov 23 23:46:38 2010
    State : active, degraded
    Active Devices : 1
    Working Devices : 1
    Failed Devices : 0
    Spare Devices : 0

    Name : WOPR.home.swankcr.com:0 (local to host WOPR.home.swankcr.com)
    UUID : e7ef8a3f:19bf62ca:86f0b395:2e58019a
    Events : 86510

    Number Major Minor RaidDevice State
    0 0 0 0 removed
    1 8 17 1 active sync /dev/sdb1


    The important parts of dmesg:
    [ 2.361475] scsi3 : sata_sil
    [ 2.361653] ata1: SATA max UDMA/100 mmio m1024@0xf8003000 tf 0xf8003080 irq 17
    [ 2.361663] ata2: SATA max UDMA/100 mmio m1024@0xf8003000 tf 0xf80030c0 irq 17
    [ 2.361671] ata3: SATA max UDMA/100 mmio m1024@0xf8003000 tf 0xf8003280 irq 17
    [ 2.361679] ata4: SATA max UDMA/100 mmio m1024@0xf8003000 tf 0xf80032c0 irq 17
    [ 2.361802] pata_amd 0000:00:07.1: version 0.4.1
    [ 2.400425] scsi4 : pata_amd
    [ 2.400705] scsi5 : pata_amd
    [ 2.400831] ata5: PATA max UDMA/100 cmd 0x1f0 ctl 0x3f6 bmdma 0xf000 irq 14
    [ 2.400839] ata6: PATA max UDMA/100 cmd 0x170 ctl 0x376 bmdma 0xf008 irq 15
    [ 2.401091] aic7xxx 0000:00:0d.0: PCI INT A -> GSI 16 (level, low) -> IRQ 16
    [ 2.710051] ata1: SATA link up 1.5 Gbps (SStatus 113 SControl 310)
    [ 2.713683] ata1.00: ATA-8: WDC WD20EARS-00MVWB0, 50.0AB50, max UDMA/133
    [ 2.713692] ata1.00: 3907029168 sectors, multi 16: LBA48 NCQ (depth 0/32)
    [ 2.732750] ata1.00: configured for UDMA/100
    [ 2.733085] scsi 0:0:0:0: Direct-Access ATA WDC WD20EARS-00M 50.0 PQ: 0 ANSI: 5
    [ 2.734123] sd 0:0:0:0: [sda] 3907029168 512-byte logical blocks: (2.00 TB/1.81 TiB)
    [ 2.734277] sd 0:0:0:0: [sda] Write Protect is off
    [ 2.734287] sd 0:0:0:0: Attached scsi generic sg0 type 0
    [ 2.734299] sd 0:0:0:0: [sda] Mode Sense: 00 3a 00 00
    [ 2.734366] sd 0:0:0:0: [sda] Write cache: enabled, read cache: enabled, doesn't support DPO or FUA
    [ 2.734970] sda: sda1
    [ 2.742665] sd 0:0:0:0: [sda] Attached SCSI disk
    [ 3.039050] ata2: SATA link up 1.5 Gbps (SStatus 113 SControl 310)
    [ 3.058888] ata2.00: ATA-8: WDC WD20EARS-00MVWB0, 50.0AB50, max UDMA/133
    [ 3.058898] ata2.00: 3907029168 sectors, multi 16: LBA48 NCQ (depth 0/32)
    [ 3.077919] ata2.00: configured for UDMA/100
    [ 3.078248] scsi 1:0:0:0: Direct-Access ATA WDC WD20EARS-00M 50.0 PQ: 0 ANSI: 5
    [ 3.078692] sd 1:0:0:0: [sdb] 3907029168 512-byte logical blocks: (2.00 TB/1.81 TiB)
    [ 3.078811] sd 1:0:0:0: Attached scsi generic sg1 type 0
    [ 3.078844] sd 1:0:0:0: [sdb] Write Protect is off
    [ 3.078851] sd 1:0:0:0: [sdb] Mode Sense: 00 3a 00 00
    [ 3.078913] sd 1:0:0:0: [sdb] Write cache: enabled, read cache: enabled, doesn't support DPO or FUA
    [ 3.079517] sdb: sdb1
    [ 3.107991] sd 1:0:0:0: [sdb] Attached SCSI disk
    [ 3.383035] ata3: SATA link down (SStatus 0 SControl 310)
    [ 3.688028] ata4: SATA link down (SStatus 0 SControl 310)
    [ 17.612045] scsi6 : Adaptec AIC7XXX EISA/VLB/PCI SCSI HBA DRIVER, Rev 7.0
    [ 17.612049] <Adaptec aic7899 Ultra160 SCSI adapter>
    [ 17.612052] aic7899: Ultra160 Wide Channel A, SCSI Id=7, 32/253 SCBs
    [ 17.612055]
    [ 17.612550] aic7xxx 0000:00:0d.1: PCI INT B -> GSI 17 (level, low) -> IRQ 17
    [ 17.648480] scsi 6:0:0:0: Direct-Access HITACHI DK32DJ-36MW AAAA PQ: 0 ANSI: 3
    [ 17.648498] scsi6:A:0:0: Tagged Queuing enabled. Depth 4
    [ 17.648533] scsi target6:0:0: Beginning Domain Validation
    [ 17.664255] scsi target6:0:0: wide asynchronous
    [ 17.676306] scsi target6:0:0: FAST-80 WIDE SCSI 160.0 MB/s DT (12.5 ns, offset 100)
    [ 17.696689] scsi target6:0:0: Ending Domain Validation
    [ 21.296726] sd 6:0:0:0: Attached scsi generic sg2 type 0
    [ 21.298104] sd 6:0:0:0: [sdc] 72205440 512-byte logical blocks: (36.9 GB/34.4 GiB)
    [ 21.306981] sd 6:0:0:0: [sdc] Write Protect is off
    [ 21.306990] sd 6:0:0:0: [sdc] Mode Sense: a7 00 10 08
    [ 21.311896] sd 6:0:0:0: [sdc] Write cache: enabled, read cache: enabled, supports DPO and FUA
    [ 21.327309] sdc: sdc1 sdc2
    [ 21.348530] sd 6:0:0:0: [sdc] Attached SCSI disk
    [ 32.823045] scsi7 : Adaptec AIC7XXX EISA/VLB/PCI SCSI HBA DRIVER, Rev 7.0
    [ 32.823050] <Adaptec aic7899 Ultra160 SCSI adapter>
    [ 32.823053] aic7899: Ultra160 Wide Channel B, SCSI Id=7, 32/253 SCBs
    [ 32.823056]
    [ 33.113490] scsi: waiting for bus probes to complete ...
    [ 36.701967] dracut: Scanning devices sdc2 for LVM logical volumes vg_wopr/root_lv vg_wopr/swaplv
    [ 36.786789] dracut: inactive '/dev/vg_wopr/root_lv' [9.78 GiB] inherit
    [ 36.786978] dracut: inactive '/dev/vg_wopr/swaplv' [2.00 GiB] inherit
    [ 37.425061] EXT3-fs: barriers not enabled
    [ 37.433841] kjournald starting. Commit interval 5 seconds
    [ 37.433898] EXT3-fs (dm-0): mounted filesystem with ordered data mode
    [ 37.473969] dracut: Mounted root filesystem /dev/mapper/vg_wopr-root_lv
    [ 37.521992] dracut: Loading SELinux policy
    [ 37.822285] SELinux: Disabled at runtime.
    [ 37.822345] SELinux: Unregistering netfilter hooks
    [ 37.826060] type=1404 audit(1290555993.738:2): selinux=0 auid=4294967295 ses=4294967295
    [ 37.881976] dracut: /sbin/load_policy: Can't load policy file /etc/selinux/targeted/policy/policy.15: No such file or directory
    [ 37.998201] dracut: Switching root
    [ 38.942687] udev[342]: starting version 161
    [ 39.609593] md: bind<sdb1>
    [ 39.998410] parport_pc 00:08: reported by Plug and Play ACPI
    [ 39.998469] parport0: PC-style at 0x378, irq 7 [PCSPP]
    [ 40.134150] ppdev: user-space parallel port driver
    [ 40.145043] 3c59x 0000:00:0f.0: PCI INT A -> GSI 18 (level, low) -> IRQ 18
    [ 40.145110] 3c59x: Donald Becker and others.
    [ 40.145131] 0000:00:0f.0: 3Com PCI 3c980C Python-T at f8880400.
    [ 40.283391] md: raid1 personality registered for level 1
    [ 40.284089] md/raid1:md0: active with 1 out of 2 mirrors
    [ 40.287417] md0: bitmap initialized from disk: read 1/1 pages, set 1 bits
    [ 40.287424] created bitmap (15 pages) for device md0
    [ 40.313751] md0: detected capacity change from 0 to 2000396697600
    [ 40.318044] md0: detected capacity change fro


    fdisk -l

    Disk /dev/sda: 2000.4 GB, 2000398934016 bytes
    81 heads, 63 sectors/track, 765633 cylinders, total 3907029168 sectors
    Units = sectors of 1 * 512 = 512 bytes
    Sector size (logical/physical): 512 bytes / 512 bytes
    I/O size (minimum/optimal): 512 bytes / 512 bytes
    Disk identifier: 0x000e9555

    Device Boot Start End Blocks Id System
    /dev/sda1 2048 3907029167 1953513560 fd Linux raid autodetect

    Disk /dev/sdb: 2000.4 GB, 2000398934016 bytes
    81 heads, 63 sectors/track, 765633 cylinders, total 3907029168 sectors
    Units = sectors of 1 * 512 = 512 bytes
    Sector size (logical/physical): 512 bytes / 512 bytes
    I/O size (minimum/optimal): 512 bytes / 512 bytes
    Disk identifier: 0x0007f811

    Device Boot Start End Blocks Id System
    /dev/sdb1 2048 3907029167 1953513560 fd Linux raid autodetect

    Disk /dev/sdc: 37.0 GB, 36969185280 bytes
    255 heads, 63 sectors/track, 4494 cylinders, total 72205440 sectors
    Units = sectors of 1 * 512 = 512 bytes
    Sector size (logical/physical): 512 bytes / 512 bytes
    I/O size (minimum/optimal): 512 bytes / 512 bytes
    Disk identifier: 0x00084d48

    Device Boot Start End Blocks Id System
    /dev/sdc1 * 2048 2099199 1048576 83 Linux
    /dev/sdc2 2099200 72204287 35052544 8e Linux LVM


    Re-adding the disk:
    mdadm --add /dev/md0 /dev/sda1
    mdadm: re-added /dev/sda1
    Personalities : [raid1]
    md0 : active raid1 sda1[0] sdb1[1]
    1953512400 blocks super 1.2 [2/1] [_U]
    [>....................] recovery = 0.0% (0/1953512400) finish=2034908.7min speed=0K/sec
    bitmap: 2/15 pages [8KB], 65536KB chunk

    unused devices: <none>


    Then, 10 seconds later:
    cat /proc/mdstat
    Personalities : [raid1]
    md0 : active raid1 sda1[0] sdb1[1]
    1953512400 blocks super 1.2 [2/2] [UU]
    bitmap: 0/15 pages [0KB], 65536KB chunk

    unused devices: <none>


    Does anyone have any ideas? I mean, once the drive is re-added it seems to work perfectly. This whole time, I've had data on the array (not something I don't have backups of), and it's 100% in tact. It's almost like the disks aren't importing by the time the array starts. It just grabs one, and runs with it.

  2. #2
    Join Date
    May 2004
    Location
    NJ
    Posts
    913

    Re: Fedora 14 - Software RAID issue

    It's probably not a software RAID issue. It's more likely an issue with WD non-RE (RAID Edition) drives. WD says that their error recovery can take up to 2 minutes to complete, causing drives without TLER (Time Limited Error Recovery) to drop out of RAID arrays.

    There is a utility you can get from WD tech support - or here - to enable and disable TLER on drives. I'd start with that.

  3. #3
    Join Date
    May 2009
    Posts
    15

    Re: Fedora 14 - Software RAID issue

    Thanks for the info, I've done some reading and will give it a shot. However, on attempting to download it from that site it's forcing me to register. And refuses to accept gmail addresses. So, I'll try to find it elsewhere.

    Boo to WD, though --- for having a linux version!

  4. #4
    Join Date
    May 2009
    Posts
    15

    Re: Fedora 14 - Software RAID issue

    I tried to use the TLER utility -- but it says the drives don't support read/write TLER.

    I contacted WD, and they just told me that I should have bought the RAID edition drives.

  5. #5
    Join Date
    May 2009
    Posts
    15

    Re: Fedora 14 - Software RAID issue

    Fedora 13 solved my issue. Not sure what's going on, but I found a few other posts on fedoraforum.org complaining about the same problem with software raid.

  6. #6
    Join Date
    Dec 2010
    Posts
    1

    Re: Fedora 14 - Software RAID issue

    I had the same issue with a Raid 5 array with 6 WD 1.5Gb EARS drives. One or more drives was slow to start up after reboot. Also TLER didn't work on my drives I believe b/c they are newer.

    I ended up delaying the assembly of the array a couple min after the machine started up.

    I edited the /etc/rc.d/rc/local file by Sleeping for 30 seconds stopping the array and then Assembling the array.

    You might want to tune the array too. I gaind 50% increase in performance by modifying the R/W cache sizes. At the recommendation of a friend.

    I was also going to look in to modifyin the ammount of time Linux will await before marking a drive as missing. That is for another day.

    -----------------------------
    touch /var/lock/subsys/local

    sleep 30
    mdadm --manage --stop /dev/md0
    sleep 5
    mdadm --assemble /dev/md0 --auto=md /dev/sd[b-f]1
    #mdadm --assemble /dev/md1 -amd /dev/sda1 /dev/sdb1 /dev/sdc1 /dev/sdd1 /dev/sdf1
    ---------------------------------------------------

  7. #7
    Join Date
    Aug 2009
    Posts
    8,475

    Re: Fedora 14 - Software RAID issue

    I was just reading a bugzilla report that looks close to this same problem.

    https://bugzilla.redhat.com/show_bug.cgi?id=649038

    interesting was down towards the end they seem to have found a workaround for the problem they were having. What it looked like was happening in this case was that udev was multi-threading the device detection on the drives and was assigning them different device names on each boot, which normally shouldn't be any problem, but it seemed that mdadm was having a little problem with it. (they say the new version of mdadm in the updates-testing repo seems to have corrected this)

    There was one that used a workaround that he said worked and that was to edit /sbin/start_udev and change

    /sbin/udevd -d

    into

    /sbin/udevd -d --children-max=1

    This caused udev to not multi-thread the device detection and so gave more cinsistant device assignments.

    Also, there is a kernel parameter that is being added now that wasn't before. in your /boot/grub/grub.conf file, you may want to check to be sure the rd_NO_MD parameter is not in there.

Similar Threads

  1. Issue with my software raid
    By migsutu in forum Servers & Networking
    Replies: 8
    Last Post: 11th December 2009, 10:56 PM
  2. Software RAID 5 with Fedora 9
    By lip in forum Using Fedora
    Replies: 1
    Last Post: 31st March 2009, 01:04 PM
  3. Software Raid Fedora 8
    By joneskent84 in forum Using Fedora
    Replies: 0
    Last Post: 13th November 2007, 10:51 PM
  4. Grub error 21. Probably software raid issue?
    By narcissist01 in forum Installation, Upgrades and Live Media
    Replies: 1
    Last Post: 30th May 2007, 08:40 PM
  5. Software Raid Fedora 4
    By Wotan in forum Installation, Upgrades and Live Media
    Replies: 1
    Last Post: 4th July 2005, 09:24 AM

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •