RHEL and Multipath

[Kaukonen]

Hello,

I got a few Red Hat Enterpice Linux running on Blade 460c g6 servers. They boot from same EVA 4400, but they are in different Blade enclosures.
I haven't done the Blade enclosure or EVA installations my self.

Servers on the other enclosure sees EVA from 4 paths.

mpath3 (3600143800648ac570000500000730000) dm-3 HP,HSV300
[size=300G][features=1 queue_if_no_path][hwhandler=0][rw]
\_ round-robin 0 [prio=100][active]
\_ 0:0:0:4 sdd 8:48 [active][ready]
\_ 1:0:0:4 sdl 8:176 [active][ready]
\_ round-robin 0 [prio=20][enabled]
\_ 0:0:1:4 sdh 8:112 [active][ready]
\_ 1:0:1:4 sdp 8:240 [active][ready]

And from the other only two
mpath0 (3600143800648ac570000500001120000) dm-0 HP,HSV300
[size=20G][features=1 queue_if_no_path][hwhandler=0][rw]
\_ round-robin 0 [prio=100][active]
\_ 0:0:0:1 sda 8:0 [active][ready]
\_ 1:0:0:1 sdb 8:16 [active][ready]

Difference in hardware is that the later has Virtual Connect and the first one connects directly to FC switches.

Is this normal "feature" of Virtual Connect or is there something wrong with my environment?

[Matti_Kurkela]

You might ask someone who knows how zoning is *supposed* to work in your FC environment.

I could be wrong, but I think this means someone has forgotten to allow your VC enclosure to see the second (stand-by?) controller of the EVA. If your EVA has a controller failure, the first enclosure could automatically switch to the other controller, but the VC enclosure would then be left without access to storage.

MK

[Kaukonen]

I asked about the zoning and they told me that they disabled two ports that have been connected cross between the FC Switches. And that is how the VC should be connected to the EVA.

This is mostlikey why there is only two channels. I have no idea if this is redudant or not.

[Matti_Kurkela · Accepted solution]

"They" might be right. It's hard to know for sure without any further information.

We don't have EVA storage systems, but I understand there is an "evainfo" tool that should help in verifying the connectivity to EVA. It can identify the EVA controller and port IDs associated with each /dev/sd* device.

Please see these threads:
https://forums.itrc.hp.com/service/forums/questionanswer.do?threadId=1432579
(download link, free HP Passport registration required)

http://forums.itrc.hp.com/service/forums/questionanswer.do?threadId=1173146
(usage example)

Unfortunately, there seems to be no separately-packaged evainfo for Linux, but it should be available in the EVA Command View media kit. Your storage admins should have such a media kit: it might be easiest to ask them to give you the Linux version of the evainfo tool.

If the evainfo tool is not available, it might be possible to dig out the equivalent information by using generic Linux tools like sdparm. Some reading of EVA documentation will probably be required too (you have to find out how EVA encodes the controller & port identifiers to the the SCSI attributes of the LUN.)

With the information produced by the evainfo tool, you could easily produce informative problem reports for your SAN admins, for example "My system X sees the LUNs through EVA's controller B only: there is no connectivity from any HBA of system X to controller A".

MK