Attached RAID, I must have misconfigured something. - SGI

This is a discussion on Attached RAID, I must have misconfigured something. - SGI ; I have an O200, running 6.5.22f. I have a SCSI RAID array attached, that's been upgraded, by replacing the drives in the array, now there's 12x250 gigs. I was told that there's a 2TB limit on a SCSI file system, ...

+ Reply to Thread
Results 1 to 2 of 2

Thread: Attached RAID, I must have misconfigured something.

  1. Attached RAID, I must have misconfigured something.

    I have an O200, running 6.5.22f. I have a SCSI RAID array attached,
    that's been upgraded, by replacing the drives in the array, now
    there's 12x250 gigs. I was told that there's a 2TB limit on a SCSI
    file system, so I've divided the array in 2(6 drives each, RAID 5).
    I'm not sure, but somewhere I must have something configured wrong. It
    seems like when I use both drives I get a lot of problems, which makes
    me think it's related to luns or something. I am getting a lot of xfs
    errors:

    Feb 1 02:59:14 1Aliver unix: ALERT: xfs_stack_trace[0]
    0xc0000000003061d0
    Feb 1 02:59:14 1Aliver unix: ALERT: xfs_da_do_buf: bno 8388608
    Feb 1 02:59:14 1Aliver unix: ALERT: dir: inode 1946159775
    Feb 1 02:59:14 1Aliver unix: ALERT: Filesystem "/backup": XFS
    internal error xfs_da_do_buf(1) at line 2176 of file ../fs/xfs
    /xfs_da_btree.c. Caller 0xc000000000328480

    I have 2 luns mapped. Now this seems wrong to me, when I fire up fx
    and I seem to succesfully see drives on

    dksc(6,1,0)
    dksc(6,1,1)
    dksc(6,2,0)
    dksc(6,2,1)

    hinv comes back with:

    Integral SCSI controller 6: Version QL1040B (rev. 2), single ended
    Disk drive: unit 1 on SCSI controller 6
    Disk drive: unit 1, lun 1 on SCSI controller 6
    Disk drive: unit 2 on SCSI controller 6
    Disk drive: unit 2, lun 1 on SCSI controller 6
    Disk drive: unit 3 on SCSI controller 6
    Disk drive: unit 3, lun 1 on SCSI controller 6
    Disk drive: unit 4 on SCSI controller 6
    Disk drive: unit 4, lun 1 on SCSI controller 6
    Disk drive: unit 5 on SCSI controller 6
    Disk drive: unit 5, lun 1 on SCSI controller 6
    Disk drive: unit 6 on SCSI controller 6
    Disk drive: unit 6, lun 1 on SCSI controller 6
    Disk drive: unit 7 on SCSI controller 6
    Disk drive: unit 7, lun 1 on SCSI controller 6
    Disk drive: unit 8 on SCSI controller 6
    Disk drive: unit 8, lun 1 on SCSI controller 6
    Disk drive: unit 9 on SCSI controller 6
    Disk drive: unit 9, lun 1 on SCSI controller 6
    Disk drive: unit 10 on SCSI controller 6
    Disk drive: unit 10, lun 1 on SCSI controller 6
    Disk drive: unit 11 on SCSI controller 6
    Disk drive: unit 11, lun 1 on SCSI controller 6
    Disk drive: unit 12 on SCSI controller 6
    Disk drive: unit 12, lun 1 on SCSI controller 6
    Disk drive: unit 13 on SCSI controller 6
    Disk drive: unit 13, lun 1 on SCSI controller 6
    Disk drive: unit 14 on SCSI controller 6
    Disk drive: unit 14, lun 1 on SCSI controller 6
    Disk drive: unit 15 on SCSI controller 6
    Disk drive: unit 15, lun 1 on SCSI controller 6

  2. Re: Attached RAID, I must have misconfigured something.

    Derrick MacPherson wrote:
    > I have an O200, running 6.5.22f. I have a SCSI RAID array attached,
    > that's been upgraded, by replacing the drives in the array, now
    > there's 12x250 gigs. I was told that there's a 2TB limit on a SCSI
    > file system, so I've divided the array in 2(6 drives each, RAID 5).
    > I'm not sure, but somewhere I must have something configured wrong. It
    > seems like when I use both drives I get a lot of problems, which makes
    > me think it's related to luns or something. I am getting a lot of xfs
    > errors:
    >
    > Feb 1 02:59:14 1Aliver unix: ALERT: xfs_stack_trace[0]
    > 0xc0000000003061d0
    > Feb 1 02:59:14 1Aliver unix: ALERT: xfs_da_do_buf: bno 8388608
    > Feb 1 02:59:14 1Aliver unix: ALERT: dir: inode 1946159775
    > Feb 1 02:59:14 1Aliver unix: ALERT: Filesystem "/backup": XFS
    > internal error xfs_da_do_buf(1) at line 2176 of file ../fs/xfs
    > /xfs_da_btree.c. Caller 0xc000000000328480
    >
    > I have 2 luns mapped. Now this seems wrong to me, when I fire up fx
    > and I seem to succesfully see drives on
    >
    > dksc(6,1,0)
    > dksc(6,1,1)
    > dksc(6,2,0)
    > dksc(6,2,1)
    >
    > hinv comes back with:
    >
    > Integral SCSI controller 6: Version QL1040B (rev. 2), single ended
    > Disk drive: unit 1 on SCSI controller 6
    > Disk drive: unit 1, lun 1 on SCSI controller 6
    > Disk drive: unit 2 on SCSI controller 6
    > Disk drive: unit 2, lun 1 on SCSI controller 6
    > Disk drive: unit 3 on SCSI controller 6
    > Disk drive: unit 3, lun 1 on SCSI controller 6
    > Disk drive: unit 4 on SCSI controller 6
    > Disk drive: unit 4, lun 1 on SCSI controller 6
    > Disk drive: unit 5 on SCSI controller 6
    > Disk drive: unit 5, lun 1 on SCSI controller 6
    > Disk drive: unit 6 on SCSI controller 6
    > Disk drive: unit 6, lun 1 on SCSI controller 6
    > Disk drive: unit 7 on SCSI controller 6
    > Disk drive: unit 7, lun 1 on SCSI controller 6
    > Disk drive: unit 8 on SCSI controller 6
    > Disk drive: unit 8, lun 1 on SCSI controller 6
    > Disk drive: unit 9 on SCSI controller 6
    > Disk drive: unit 9, lun 1 on SCSI controller 6
    > Disk drive: unit 10 on SCSI controller 6
    > Disk drive: unit 10, lun 1 on SCSI controller 6
    > Disk drive: unit 11 on SCSI controller 6
    > Disk drive: unit 11, lun 1 on SCSI controller 6
    > Disk drive: unit 12 on SCSI controller 6
    > Disk drive: unit 12, lun 1 on SCSI controller 6
    > Disk drive: unit 13 on SCSI controller 6
    > Disk drive: unit 13, lun 1 on SCSI controller 6
    > Disk drive: unit 14 on SCSI controller 6
    > Disk drive: unit 14, lun 1 on SCSI controller 6
    > Disk drive: unit 15 on SCSI controller 6
    > Disk drive: unit 15, lun 1 on SCSI controller 6



    The only time I have seen something like this where the controller
    thinks it has 16 drives is when you have the drive set at the same
    address as the controller.

    SGI by default use ID 0 for the controller. Is your raid set for 0 as well.

    Rob


+ Reply to Thread