BladeSystem - General
1745797 Members
3679 Online
108722 Solutions
New Discussion юеВ

p700m, 3G SAS switches, MDS600 - lost redundant path

 
John T Willis
Advisor

p700m, 3G SAS switches, MDS600 - lost redundant path


c7000 chassis with blades and two 3G SAS switches, MDS600 attached to both 3G SAS switches

in a redundant configuration

HP recently replaced one drawer in the MDS600

ever since the ACU reports a Warning Alert

799 Logical Drive 1 (20.0 TB, RAID 1+0) is degraded due to the loss of a redundant path

the bay 1 drive in each drawer now reports under the Physical view

Transfer Speed: PHY 1: Unknown, PHY 2: 3.0 GBPS

Location:
Port 52
Box 1
Bay 1

No pathing information at all.

34 other drives in each drawer report

Transfer Speed: PHY 1: 3.0GBPS, PHY 2: 3.0 GBPS

Location:
Port 52
Box 1
Bay 2
Active Path
Port 52: Box 1 : Bay 2 - OK
Redundant Path(s)
Port 62: Box 1 : Bay 2 - OK

I believe this rules out bad controller, bad switches, bad i/o modules

Could the SAS topology software have a bug?

Firmware as follows:

3G SAS - 2.2.15
MDS600 - 2.66

Plan to engage with HP on Monday regarding the missing information, but since this is perhaps a software issue, not sure how to approach the hardware channel for support.
4 REPLIES 4
Johan Guldmyr
Honored Contributor

Re: p700m, 3G SAS switches, MDS600 - lost redundant path

As for how to approach support: sometimes it works to let them come to the same conclusion as you have. In either case the firmware of the systems were handled by the same team as the hardware guys. I presume it's still this way.

Interesting problem though.

Restarted?
John T Willis
Advisor

Re: p700m, 3G SAS switches, MDS600 - lost redundant path

contacted support, they collected a Diagnostic and an HPS report and are studying the problem

your right it wasn't that difficult, they kicked it up to L2 support right away

John T Willis
Advisor

Re: p700m, 3G SAS switches, MDS600 - lost redundant path

 

tried re-flashing the mds600 firmware, no effect

 

L2 support took a look at our Visio diagram of our SAS cabling and determined it was not a supported configuration.

 

It has been working this way for over 7 months, but its not supported so we have to schedule down time to re-cable, then if the problem doesn't go away I assume we'll pick up the problem from there.

 

in the mean time HP contacted us to verify all our mds600 units had been recalled, so before recabling we're looking at scheduling a larger window to replace all of the mds600 units.

 

problem is best practice was to located them at the bottom of "full" 42 unit racks.. so we're a little concerned that some of the over weight may make getting the old mds600 units out of the rack difficult.. bending and twisting due to over weight can make the bottom of the racks hard to deal with.

 

its going to be a long summer..

John T Willis
Advisor

Re: p700m, 3G SAS switches, MDS600 - lost redundant path

 

just a note.

 

the mds600 replacements arrived at our data center.

 

on Monday we're holding a meeting with staff to scheduling replacing the units.

 

once they are replaced we'll either resume working the redundancy problem with L2 support or the problem will be resolved.

 

some discussion will be had about whether to bundle the re-cabling event with the recall replacement event