SLES10 SP3 multipath failed

[Sambugaro Fabio]

Hello,

we installed SLES 10 SP3. I configure the server with multipath to use a EVA4400 LUN. To test the path I disable a port in the fibre channel switch and I see the following situation after command multipath -ll:

mpathb (3600508b4000c212d0000800006530000) dm-0 HP,HSV300
[size=20G][features=1 queue_if_no_path][hwhandler=0]
\_ round-robin 0 [prio=1][active]
\_ 1:0:1:1 sdd 8:48 [active][ready]
\_ round-robin 0 [prio=1][enabled]
\_ 1:0:0:1 sdc 8:32 [active][ready]
\_ round-robin 0 [prio=0][enabled]
\_ 0:0:1:1 sdb 8:16 [failed][faulty]
\_ round-robin 0 [prio=0][enabled]
\_ 0:0:0:1 sda 8:0 [failed][faulty]

Than we enable the fc port again but the 2 paths were in the same status. Why? The state should be [active][Ready] again. Any idea?
Thanks in advance.

Fabio Sambugaro

[Alzhy]

Greetings,

First thing first - do you even have the correct HP Device Mapper Multipath Enablement Kit instaleld for your version of SUSE Linux?

You should have HPDM installed sir as well as suggested settings for your HBA (/etc/modprobe.conf)

With HPDMin place, proper settings for EVA are automatically in /etc/multipath.conf


Cheers.

[Sambugaro Fabio]

Hello,

first of all, my hardware is Bl25p G0. From SLES10 SP3 and for SLES11 you do not need to install the Device Mapper Multipath Enablement Kit see link
http://h20000.www2.hp.com/bizsupport/TechSupport/SoftwareDescription.jsp?lang=en&cc=us&prodTypeId=12169&prodSeriesId=3559651&swItem=co-82195-1&prodNameId=3559652&swEnvOID=2078&swLang=13&taskId=135&mode=4&idx=0

in the overview you find: This is the final release of HP Device Mapper Multipath Enablement Kit. A new Native Linux Device Mapper Multipath reference guide has been put together. User can follow the new guide to setup linux Device Mapper Multipath for HP arrays. The guide will be updated as new arrays are released and with other relevant up-to-date information. The Native Linux Device Mapper Multipath reference guide can be found on SPOCK both internally and externally at http://www.hp.com/storage/spock. On the web site, look for â Application Notesâ on left hand side and click on â Solutions: Linuxâ

So you should use the multipath in-box on SuSE. For the qla driver is the same. The only software that you must install from Hp is the hp-fc-enablement kit, see link
http://h20000.www2.hp.com/bizsupport/TechSupport/SoftwareDescription.jsp?lang=en&cc=us&prodTypeId=3709945&prodSeriesId=1120361&swItem=co-66214-1&prodNameId=451433&swEnvOID=2078&swLang=13&taskId=135&mode=4&idx=0

choose release notes, operating system section: ...NOTE: Starting with RHEL 5 U3, SLES 10 SP3, and SLES 11, Fibre Channel HBAs and mezzanine cards are supported by Red Hat and Novell in-box drivers (included in the OS distribution), and multipath failover is handled by Device Mapper...

However, for the /etc/modprobe.conf.local and the /etc/multipath.conf I follow the suggestion given in the Device Mapper Multipath Enablement Kit and the Linux Driver Kit for HP Qlogic HBA release notes.
Any suggestion.
Thanks in advance.

Fabio

[Alzhy]

Hmm strange then.

Have you tried re-scanning your FC HBAs (use your fibreutil tools or echo the fc hosts)? Also have you done a multipath -vN (2 or 1 or 0) ?

I have EVAs in my Linux ecosystems and fallback works like a charm.

You may also check on the VDISK properties on the EVA too -- if VDISK is set to fall back.

[Sambugaro Fabio]

Hello,

what is the version of your operating system. All my server run on SLES but only this with SLES10 SP3 and qlogic in this server type. This is the first time I use the multipath features.
Thanks.

Fabio

[Tim Nelson]

two quick thoughts.. after re-enabling the port are you executing multipath or multipath -ll ?

also, is your multipathd daemon running ?
service multipathd status

[Sambugaro Fabio]

Hi,

yes I executed the multipath and multipath -ll command, but the physical path remains in state "faulty", and the daemon is always running. I also try to rescan all the path, but the result is the same.
Thanks.

Fabio

[Reto Kisseleff]

Hi Fabio
you're right you "should" use Linux inbox driver if your running on SLES10 SP3 or RHEL5 U3 (or higher in both cases). However what you need is the right multipath.conf file, to let multipath know how it should do the job with different storage types.
Therefore I'd recommend to use the config file from HPDM. I'll attach you a copy for SLES10. (if you have problems with multipath, the config file can be interesting to look at
cheers
Reto

[Sambugaro Fabio]

Hi,

the multipath.conf is configured as you suggest:

defaults {
udev_dir /dev
polling_interval 10
selector "round-robin 0"
path_grouping_policy failover
getuid_callout "/sbin/scsi_id -g -u -s /block/%n"
prio const
path_checker tur
rr_min_io 100
rr_weight uniform
failback immediate
no_path_retry 12
user_friendly_names yes
}

#For EVA4x00/EVA6x00/EVA8x00
device {
vendor "HP|COMPAQ"
product "HSV1[01]1 \(C\)COMPAQ|HSV2[01]0|HSV300|HSV4[05]"
getuid_callout "/sbin/scsi_id -g -u -s /block/%n"
prio aula
hardware_handler "0"
path_selector "round-robin 0"
#path_grouping_policy group_by_prio
path_grouping_policy failover
failback immediate
rr_weight uniform
rr_min_io 100
no_path_retry 18
path_checker tur
}

But the problem still persist.
Thanks.

Fabio

[Reto Kisseleff]

Hi Fabio

yeah looks like the same, except the path_grouping_policy, but this normally don't lead to such behaviour.
Then I would recommend first to upgrade multipath to the most up2day version. Additional you should check SAN switch and HBA firmware as second step.

If that doesn't help, you should open a case to HP.

cheers
Reto