I'm trying to run multipath on a Debian Lenny box with 2 virtual disks provided by our Intel Modular Server system. (Intel officially supports SLES and RHEL only, with Debian you're left alone in the dark.) We have 2 disks with 2 paths each (sda through sdd), where /dev/sda is the Linux 'system' disk (with 2 partitions) and the 'database' disk is recogized as /dev/sdd (single partition). The other two disks (paths, really) are not accessible (executing
fdisk -l /dev/sdb /dev/sdc yields nothing).
My problem currently is that multipath recognizes the database disk only (sdb and sdd apparently, with sdd being the
active path and sdb's path state being
failed instead of the desired
enabled state). The system disk is not listed by the
multipath -ll command even though the
/var/lib/multipath/bindings file lists both devices with their correct IDs. Of course,
/dev/mapper doesn't provide symlinks to the system disk either.
Code:
~# multipath -ll
database (22209000155faaffa) dm-0 Intel ,Multi-Flex
[size=800G][features=1 queue_if_no_path][hwhandler=0]
\_ round-robin 0 [prio=2][active]
\_ 0:0:0:1 sdb 8:16 [failed][ready]
\_ 0:0:1:1 sdd 8:48 [active][ready]
Does anyone know how to get around this issue and
make the system disk show up? Below is my current
/etc/multipath.conf configuration for sake of completeness:
Code:
defaults {
user_friendly_names yes
}
blacklist {
devnode "^(ram|raw|loop|fd|md|dm-|sr|scd|st)[0-9]*"
devnode "^(hd|xvd)[a-z][[0-9]*]"
devnode "^cciss!c[0-9]d[0-9]*[p[0-9]*]"
}
devices {
device {
vendor "Intel"
product "Multi-Flex"
path_grouping_policy "group_by_prio"
getuid_callout "/lib/udev/scsi_id -g -u /dev/%n"
prio "intel"
path_checker tur
path_selector "round-robin 0"
# hardware_handler "1 alua"
failback immediate
rr_weight uniform
rr_min_io 100
no_path_retry queue
features "1 queue_if_no_path"
}
}
multipaths {
multipath {
wwid 222ef0001555ab385
alias system
}
multipath {
wwid 22209000155faaffa
alias database
}
}
Note: I have commented out the
hardware_handler option; the multipath daemon comments on the "1 alua" value with:
unknown hardware handler type
(No wonder, according to the multipath.conf manpage "1 emc" is the only implemented value!)
The configuration is based on the Intel's MPIO config procedure for SLES (SuSE Linux Enterprise Server), instead of the RPMs shipped by Intel I installed
multipath-tools using apt-get.
Anyone had this or a similar issue before and solved it?
P.S., some resources I found helpful so far:
Multipath on Debian and Intel Modular Server (MFSYS25)
[SOLVED] Multipath on Debian Lenny and Intel Modular Server
