Need help for RHEL 5.5 Multipath problem with MSA2000fx

[Aungshuman Paul]

Hi,

Can anyone help me regarding the following issue:

We installed RHEL 5.5 64 bit on a DL380 G5. And present a disk from MSA2000fc. But the problem it shows multiple disk :

[root@localhost ~]# fdisk -l

Disk /dev/cciss/c0d0: 293.5 GB, 293563949056 bytes
255 heads, 63 sectors/track, 35690 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes

Device Boot Start End Blocks Id System
/dev/cciss/c0d0p1 * 1 25 200781 83 Linux
/dev/cciss/c0d0p2 26 2065 16386300 82 Linux swap / Solaris
/dev/cciss/c0d0p3 2066 35690 270092812+ 83 Linux

Disk /dev/sda: 99.9 GB, 99999989760 bytes
255 heads, 63 sectors/track, 12157 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes

Disk /dev/sda doesn't contain a valid partition table

Disk /dev/sdb: 99.9 GB, 99999989760 bytes
255 heads, 63 sectors/track, 12157 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes

Disk /dev/sdb doesn't contain a valid partition table

Disk /dev/sdc: 99.9 GB, 99999989760 bytes
255 heads, 63 sectors/track, 12157 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes

Disk /dev/sdc doesn't contain a valid partition table

Disk /dev/sdd: 99.9 GB, 99999989760 bytes
255 heads, 63 sectors/track, 12157 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes

Disk /dev/sdd doesn't contain a valid partition table

Waiting for valuable reply.

Aungshu

[Matti_Kurkela]

This is called "multipathing" and it's normal in fault-tolerant FC storage installations.

Instead of one /dev/sd* device per LUN, you'll get one /dev/sd* device for each _path_ the LUN can be accessed through. If you have (for example) two FibreChannel HBAs on your server, one FC switch and two connections to the MSA2000fc, here's how you get 4 paths:
- from HBA 1 to switch to MSA2000fc port A
- from HBA 2 to switch to MSA2000fc port A
- from HBA 1 to switch to MSA2000fc port B
- from HBA 2 to switch to MSA2000fc port B

If you configure your server to support multipathing, your server can keep accessing the disk even if some connections are lost, as long as there is one good connection from the server to the switch and from the switch to the MSA.

RedHat recommends using the RHEL 5.x's built-in DM-Multipathing. Please see:

http://docs.redhat.com/docs/en-US/Red_Hat_Enterprise_Linux/5/html/DM_Multipath/index.html

MK

[Alzhy]

Aungshurman,

It is advisable that you install the HP Device Multipather Enablement Kit. Since you are using RHEL 5.5, you need HPDM 4.4.1.

Google HP Device Mapper 4.4.1 or check this link: http://bizsupport1.austin.hp.com/bizsupport/TechSupport/SoftwareDescription.jsp?lang=en&cc=us&prodTypeId=18964&prodSeriesId=3559651&prodNameId=3559652&swEnvOID=2027&swLang=13&mode=2&taskId=135&swItem=co-82195-1

With HPDM, you'll get /etc/multipath.conf "tuned" for whatecer flavour of HP StorageWorks array you may have -- incluing the MSA you described.

Once it is installed and your system rebooted, your MSA storage will now have /dev/mapper/mpathNN entries - whose "device paths" you can check via "multipath -ll".

This device mapper entry /dev/mapper/mpathNN is now what you should fdisk and partition and USE for whatever.

You may also want to use friendly names by having multipather "aliases" in /etc/multipath.conf so your device mapper names can have much meaningful names like: /dev/mapper/ASMDISK01.

Pls. be familar with HP Device Mapper and Linux Multipath -- it is really simple.


Hope this helps.


N

[Alzhy]

BTW, here's a snippet of /etc/multipath.conf with HPDM Enablement kit installed:

# The multipaths section - uncomment this section to define a per multipath
# device settings.

multipaths {

multipath {
wwid 3xxxxxxb40006a4640000a000011b0000
alias ASMDISK001
}
multipath {
wwid 3xxxxxxb40006a5d50000f00001160000
alias ASMDISK002
}

}



# For MSA23xxfc and MSA23xxi arrays
device {
vendor "HP"
product "MSA2312fc|MSA2324fc|MSA2312i|MSA2324i"
path_grouping_policy group_by_prio
getuid_callout "/sbin/scsi_id -g -u -s /block/%n"
path_checker tur
path_selector "round-robin 0"
prio_callout "/sbin/mpath_prio_alua /dev/%n"
rr_weight uniform
failback immediate
hardware_handler "0"
no_path_retry 18
rr_min_io 100
}