P6300 link to centos5.6 with multipath

[mukglion]

Hi guys

We have a P6300 ,it link to centos5.6 via san switch.Each controller has one fc cable link to san switch , and server has two fc cables linke to san switch.

When i type multipath -ll, display as blow

[root@hp1 device]# multipath -ll
mpath2 (36001438009b01ddb0000900000490000) dm-0 HP,HSV340
[size=15G][features=1 queue_if_no_path][hwhandler=0][rw]
\_ round-robin 0 [prio=100][enabled]
 \_ 1:0:1:1 sdc 8:32  [active][ready]
 \_ 2:0:1:1 sde 8:64  [active][ready]
\_ round-robin 0 [prio=20][enabled]
 \_ 1:0:0:1 sdb 8:16  [active][ready]
 \_ 2:0:0:1 sdd 8:48  [active][ready]

It`s seem like one controller is primary and another one is secondary.

The /etc/multipath.conf as blow

devices  {
        device {
                vendor                  "HP"
                product                 "HSV2[01]0|HSV3[046]0|HSV4[05]0"
                path_grouping_policy    group_by_prio
                getuid_callout          "/sbin/scsi_id -g -u -s /block/%n"
                prio_callout            "/sbin/mpath_prio_alua /dev/%n"
                path_selector           "round-robin 0"
                path_checker            readsector0
                hardware_handler        "0"
                failback                immediate
                rr_weight               uniform
                rr_min_io               100
                no_path_retry           2
        }
}

In test, we have a problem. When i pull out primary controller fc cable, I/O needs to wait about 47s, then I/O will move to another

link. The waiting time is too log for oracle.

Guys, is there any way to made waiting time be short?

[Ken Grabowski]

The HP recommended settings for EVA4x00/EVA6x00/EVA8x00/P6300/P6500 are:

device {
vendor “HP”
product "HSV2[01]0|HSV3[046]0|HSV4[05]0"
path_grouping_policy group_by_prio
getuid_callout "/sbin/scsi_id -g -u -s /block/%n"
prio_callout "/sbin/mpath_prio_alua /dev/%n"
path_selector "round-robin 0"
path_checker tur
hardware_handler "0"
failback immediate
rr_weight uniform
rr_min_io 100
no_path_retry 18

}

Try matching those settings first and if you still have a problem I would look at your Oracle io timeout options.  You may need to tune the database too.

Multipath for HP Storage Disk

[mukglion]

I will try it

Thanks, ken

[fssdc]

Hi,

In Debian squeeze multipath-tool does not ship /sbin/scsi_id nor /sbin/mpath_prio_alua.

Do you know what should be replacing these callouts ?

getuid_callout "/sbin/scsi_id -g -u -s /block/%n"
prio_callout "/sbin/mpath_prio_alua /dev/%n"

Thanks.

[Matti_Kurkela]

According to packages.debian.org, the scsi_id is located in /lib/udev/scsi_id, in package udev. It is probably used for other things besides multipath disk identification.

The mpath_prio_alua does not seem to exist in Debian at all. Instead, the multipath-tools package contains /lib/multipath/libprioalua.so, which probably serves much the same purpose as RedHat/Centos 5.x /sbin/mpath_prio_alua, but in the form of a shared library instead of an independent binary.

[fssdc]

Ok, Thanks.

So I guess I should left out "prio_callout", right ?

[Matti_Kurkela]

Read the example configuration files and other documentation available in /usr/share/doc/multipath-tools.

Using the prio_callout /sbin/mpath_prio_alua for RHEL/CentOS implies that the storage system is of active/passive type and is expected to use the ALUA standard for identifying the active storage paths. So you should use the equivalent setting, which is:

prio "alua"