[dm-devel] anyone using this setup without issues

["Kevin Foote" <kevin foote gmail com>]

Here is my setup: (this is ISCSI) going for multipath failover.

Equallogic storage
Qlogic 4052c HBAs (2 ports)
RHEL4.5
device-mapper-multipath-0.4.5-21.RHEL4
HP-GigE switches
/etc/multipath.conf      using failover mode
defaults {
        multipath_tool          "/sbin/multipath -v0"
        udev_dir                /dev
        #polling_interval       2
        polling_interval        4
        selector                "round-robin 0"
        path_grouping_policy    failover 
        getuid_callout          "/sbin/scsi_id -g -u -s /block/%n"
        path_checker            readsector0
        prio_callout            "/bin/true"
        features                "0"
        #rr_min_io               2
        rr_min_io               100
        rr_weight               priorities      
        failback                immediate
        no_path_retry           fail
        user_friendly_name      yes
}

Here is the problem:
Dec 20 05:28:32 192.168.46.31/192.168.46.31 kernel: SCSI error : <2 0 5 0> return code = 0x10000
Dec 20 05:28:32 192.168.46.31/192.168.46.31 multipathd: 65:16: readsector0 checker reports path is up
Dec 20 05:28:32 192.168.46.31/192.168.46.31
 kernel: device-mapper: dm-multipath: Failing path 8:96.
Dec 20 05:28:32 192.168.46.31/192.168.46.31 multipathd: 65:16: reinstated
Dec 20 05:28:32 
192.168.46.31/192.168.46.31 multipathd: ban1.banprod: remaining active paths: 2
Dec 20 05:28:32 192.168.46.31/192.168.46.31 multipathd: 8:96: readsector0 checker reports path is down
Dec 20 05:28:32 192.168.46.31/192.168.46.31 multipathd: checker failed path 8:96 in map ban1.datafiles.famp
Dec 20 05:28:32 
192.168.46.31/192.168.46.31 multipathd: ban1.datafiles.famp: remaining active paths: 1
Dec 20 05:28:35 192.168.46.31/192.168.46.31 kernel: SCSI error : <2 0 8 0> return code = 0x10000
Dec 20 05:28:35 192.168.46.31/192.168.46.31 multipathd: 8:144: readsector0 checker reports path is down
Dec 20 05:28:35 
192.168.46.31/192.168.46.31 multipathd: checker failed path 8:144 in map ban1.datafiles.wflw
Dec 20 05:28:35 192.168.46.31/192.168.46.31 kernel: device-mapper: dm-multipath: Failing path 8:144.
Dec 20 05:28:37 192.168.46.31/192.168.46.31 multipathd: ban1.datafiles.wflw: remaining active paths: 1
Dec 20 05:28:37 192.168.46.31/192.168.46.31
 multipathd: 8:96: readsector0 checker reports path is up
Dec 20 05:28:37 192.168.46.31/192.168.46.31 multipathd: 8:96: reinstated
Dec 20 05:28:37 
192.168.46.31/192.168.46.31 multipathd: ban1.datafiles.famp: remaining active paths: 2
Dec 20 05:28:39 192.168.46.31/192.168.46.31 multipathd: 8:144: readsector0 checker reports path is up
Dec 20 05:28:39 192.168.46.31/192.168.46.31 multipathd: 8:144: reinstated
Dec 20 05:28:39 192.168.46.31/192.168.46.31
 multipathd: ban1.datafiles.wflw: remaining active paths: 2

There is no real pattern to these disconnect/reconnect events. The only setting I'm thinking about changing 
currently is the no_path_retry. I'm thinking of making this 2 or some low number. The equallogic shows load balancing
events but they don't always align with the drop/reconnect that I'm seeing on the linux hosts.

Is anyone using these HBAs with equallogic on RHEL4.5 ??? 

-- 
:wq!
kevin.foote