[dm-devel] multipathing flip flopping on Sun 6140

[James Fillman]

I'm having a major problem with my multipathing on my RHEL5 servers.

Here's the problem:

The paths to my SAN volume keep flip flopping between the two
controllers. Sometimes I see the path with the lowest priority set as
the 'active' path with the higher priority path set to 'enabled'.
Sometimes I see both paths set to 'enabled'. dmesg is contantly
outputting path error's and the SAN keep reporting error's about the
volume is not being managed by its preferred controller.

I had multipathing working great on our STK D280 SAN with brocade
switches but we've recently upgraded to the STK 6140 SAN with qlogic
5200 switches.

Here's my setup:

- RHEL5-64bit server
- stock qla2xxx hba driver
- device-mapper-multipath
- Single Qlogic QLE2460 HBA
- Sun's 6140 SAN
- Qlogic 5200 switch

My server has a single HBA connected to the switch which then connects
to each controller.

I've learned that for this hardware configuration, multipathing needs
AVT enabled so the fibre channel switch port for my server has its
'Host' type set to 'Sun with Veritas DMP'. My multipathing config is
using 'group_by_prio' for the path grouping policy.

Here's my multipath.conf:

blacklist
{
         devnode "^(ram|raw|loop|fd|md|dm-|sr|scd|st|sda)[0-9]*"
}

defaults
{
        udev_dir                /dev
        polling_interval        10
        selector                "round-robin 0"
        path_grouping_policy    group_by_prio
        getuid_callout          "/sbin/scsi_id -g -u -s /block/%n"
        prio_callout            "/sbin/mpath_prio_tpc /dev/%n"
        path_checker            tur
        rr_min_io               100
        rr_weight               priorities
        failback                manual
        no_path_retry           fail
        user_friendly_name      yes
}

devnode_blacklist
{
        devnode "^(ram|raw|loop|fd|md|dm-|sr|scd|st|sda)[0-9]*"
        devnode "^hd[a-z][0-9]*"
        devnode "^cciss!c[0-9]d[0-9]*"
}
devices
{
        device
        {
                vendor                  "SUN"
                product                 "CSM200_R"    # the 6140 reports
that it's a CSM200_R
                path_grouping_policy    group_by_prio
                path_checker            tur
                prio_callout            "/sbin/mpath_prio_tpc /dev/%n"
                getuid_callout          "/sbin/scsi_id -g -u -s
/block/%n"
        }
}

Here's some different output from the multipath command showing the
different states of my paths:

[root at plxp02log etc]# multipath -v2 -ll
mdi_logging (3600a0b800029a482000005d34695182b) dm-8 SUN,CSM200_R
[size=100G][features=0][hwhandler=0]
\_ round-robin 0 [prio=3][enabled]
 \_ 1:0:0:1 sdc 8:32  [active][ready]
\_ round-robin 0 [prio=0][enabled]
 \_ 1:0:1:1 sdf 8:80  [active][ready]

mdi_logging (3600a0b800029a482000005d34695182b) dm-8 SUN,CSM200_R
[size=100G][features=0][hwhandler=0]
\_ round-robin 0 [prio=3][enabled]
 \_ 1:0:1:1 sdf 8:80  [active][ready]
\_ round-robin 0 [prio=0][active]
 \_ 1:0:0:1 sdc 8:32  [active][ready]

dmesg output looks like this:
device-mapper: multipath: Failing path 8:32.
device-mapper: multipath: Failing path 8:80.

It looks like the path checker keeps failing and/or the priority of each
path keeps switching.

I don't know much anything about configuring san's or fibre channel
switches. I've only started learning the in's and out's of setting up
multipathing and so I don't really know how to troubleshoot this
problem. Is my multipathing config correct for this hardware?

Like I said before, it was working great on a D280 with a brocade
switch. I don't have access to the SAN gear and my SAN guy is pointing
the finger at my server config.

If someone could validate my config, tell what I'm doing right and what
I'm doing wrong, I'd be forever grateful. I've not found much info
online pertaining to this hardware config.

thanks,
James Fillman