Tuesday, January 11, 2011

How to verify Linux fibre channel

How to verify Linux fiber channel
Bottom of Form
How to verify Linux fiber channel configurations with multipathing

How to verify Linux fibre channel configurations with multipathing I/O (MPIO).

Solution
  1. Verify that multipath has the correct number of paths to the LUN and that the status is ready.

    # multipath –ll

    mpath0 (360a9800043346441464a44424c746e54)
    [size=10 GB][features="1 queue_if_no_path"][hwhandler="0"]
    \_ round-robin 0 [prio=8][active]
    \_ 0:0:0:0 sda 8:0   [failed][faulty]
    \_ 1:0:0:0 sdi 8:128 [active][ready]
    \_ round-robin 0 [prio=2][enabled]
    \_ 0:0:1:0 sde 8:64  [active][ready]
    \_ 1:0:1:0 sdm 8:192 [active][ready]

  2. The dm-multipath device is mpath0 which is comprised of SCSI devices sde, sdm, sda and sdi. The SCSI devices are grouped by priority.  Priority 2 is used for sde and sdm and priority 8 is used for sda and sdi. The groups are ordered so that the higher value priority is used first. In this example, multipath shows that access to the primary path through device sda has failed and is labeled faulty.Locate the storage controller FCP target ports associated with the failed paths.
  3. Locate the device's host id, channel id, target id and LUN ID.

# ls -la /sys/block/sda/device
lrwxrwxrwx  1 root root 0 Aug 28 22:37 /sys/block/sda/device -> ../../devices/pci0000:00/0000:00:06.0/0000:05:00.0/0000:06:01.0/host0/target0:0:0/0:0:0:0

The ID values are interpreted left to right and represent the host, channel, target and LUN, each of which is 0.
Next locate the proc file name that is associated with the host controller.

# cat /sys/class/scsi_host/host0/proc_name
qla2xxx

# cat /sys/class/scsi_host/host0/unique_id
0

Finally display the contents of the proc file to determine the node and port names of the target using the proc_name, and unique_id entries.

# cat /proc/scsi/qla2xxx/0
. . .
SCSI Device Information:
scsi-qla0-adapter-node=200000e08b9dc0be;
scsi-qla0-adapter-port=210000e08b9dc0be;
scsi-qla0-target-0=500a09839617c3ac;
scsi-qla0-target-1=500a09838617c3ac;
The FCP target port WWPN associated with the LUN is target 0 and is highlighted above.Verify that the FCP target port is online and available to the host initiator. Once the connectivity problem is corrected, the path will return to active and ready.

  1. # multipath -ll
    mpath0 (360a9800043346441464a44424c746e54)
    [size=10 GB][features="1 queue_if_no_path"][hwhandler="0"]
    \_ round-robin 0 [prio=8][active]
    \_ 0:0:0:0 sda 8:0   [active][ready]
    \_ 1:0:0:0 sdi 8:128 [active][ready]
    \_ round-robin 0 [prio=2][enabled]
    \_ 0:0:1:0 sde 8:64  [active][ready]
    \_ 1:0:1:0 sdm 8:192 [active][ready]
  2. Troubleshooting Path Problems
  3. Link failures and fabric zoning can lead to a primary FCP target port to becoming unavailable to the host. The following steps may be used to isolate such a condition.
  4. From both storage controller nodes, determine if the FCP target ports are online using the command fcp show adapters. Verify the FCP target ports for the hosting storage controller node are in an ONLINE state.
  5. Verify the host's fibre channel storage controllers are logged in to the fabric.
  6. Determine if fabric cabling for the host and storage controller target conforms to NetApp best practices.  See the Fibre Channel and iSCSI Configuration Guide and Data ONTAP Block Access Management Guide for acceptable cabling configurations.
  7. If the host and target ports are online, then the fabric zoning configuration should verif

1 comment: