Unknown HW PATH, but everything else looks good?

Okay, not so good.  I just discovered that the underlying sd# devices for psuedodevice emcpowera are not listed under  /dev.  Under /dev, there is no sdm or sdn.
Eeek.

Hi,

It most likely means these devices are gone on the OS level - so either they changed their names after some reconfiguration, or they are gone for good. You can get the "inq" utility and see whether the LUN "vm-box SC 107" is visible down all 4 paths - if not, check the Storage Group configuration on the array, then rescan the HBAs and/or reboot the host and run "powermt config" - then check again.

Let us know how it goes...

Thanks!

Pawel

Edit:  found the inq tool..

I'll post the results in a reply

but I'm very certain the other paths are good. I think it's just that the dev labels changed on the one pair of paths (for SP A1 and B1)

Here are the inq results, underneath the powermt display output:

xen-R9B2:~ # powermt display dev=all

Pseudo name=emcpowera

VNX ID=APM00142118655 [XEN-R9B2]

Logical device ID=60060160F02039001280043534EDE511 [vm-users-SC 107]

state=alive; policy=CLAROpt; queued-IOs=0

Owner: default=SP B, current=SP B       Array failover mode: 4

==============================================================================

--------------- Host ---------------   - Stor -  -- I/O Path --   -- Stats ---

###  HW Path               I/O Paths    Interf.  Mode     State   Q-IOs Errors

==============================================================================

   7 qla2xxx                sdb         SP B4    active   alive      0      0

   7 qla2xxx                sdc         SP A4    active   alive      0      0

   8 UNKNOWN                sdm         SP A1    active   alive      0      0

   8 UNKNOWN                sdn         SP B1    active   alive      0      0

Pseudo name=emcpowerc

VNX ID=APM00142118655 [XEN-R9B2]

Logical device ID=60060160F0203900344155BBCCF1E511 [HQPO2-109-SC]

state=alive; policy=CLAROpt; queued-IOs=1

Owner: default=SP A, current=SP A       Array failover mode: 4

==============================================================================

--------------- Host ---------------   - Stor -  -- I/O Path --   -- Stats ---

###  HW Path               I/O Paths    Interf.  Mode     State   Q-IOs Errors

==============================================================================

   7 qla2xxx                sde         SP A4    active   alive      1      0

   7 qla2xxx                sdd         SP B4    active   alive      0      0

   8 qla2xxx                sdr         SP B1    active   alive      0      0

   8 qla2xxx                sdq         SP A1    active   alive      0      0

xen-R9B2:~ #

xen-R9B2:~ # ./inq.LinuxAMD64

Inquiry utility, Version V8.2.0.0 (Edit Level: 2153) built with SYMAPI Version V8.2.0.0 (Edit Level 2153)

Copyright (c) [1997-2016] EMC Corporation. All Rights Reserved.

For help type inq -h.

............

-------------------------------------------------------------------------

DEVICE         :VEND    :PROD            :REV   :SER NUM          :CAP(kb)

-------------------------------------------------------------------------

/dev/sda       :DELL    :PERC H310       :2.12  :           :   976224256

/dev/sdb       :DGC     :VRAID           :0533  :6B56EC43   :  4294967296

/dev/sdc       :DGC     :VRAID           :0533  :6B56EC43   :  4294967296

/dev/sdd       :DGC     :VRAID           :0533  :6D56F3F9   :  1073741824

/dev/sde       :DGC     :VRAID           :0533  :6D56F3F9   :  1073741824

/dev/sdf       :Dell    :Internal Dual SD:1.:   :           :     1992704

/dev/sdk       :DGC     :VRAID           :0533  :6B56EC43   :  4294967296

/dev/sdl       :DGC     :VRAID           :0533  :6B56EC43   :  4294967296

/dev/sdq       :DGC     :VRAID           :0533  :6D56F3F9   :  1073741824

/dev/sdr       :DGC     :VRAID           :0533  :6D56F3F9   :  1073741824

/dev/emcpowera :DGC     :VRAID           :0533  :6B56EC43   :  4294967296

/dev/emcpowerc :DGC     :VRAID           :0533  :6D56F3F9   :  1073741824

xen-R9B2:~ #

So yeah, there is no /dev/sdm or sdn.

Well, I've tried powermt config,  no change.

I ran "powermt remove dev= sd "  to remove the two quirky paths, then powermt config, it comes back the same.

Removed again, ran /etc/opt/emcpower/emcplun_linux scan hba,  (since we're on SLES11)  and it still comes back the same. Bizarre.

It's also weird that Powerpath reports this path as both "alive" and "active".
I really don't want to resort to a reboot, as

1) This only affects one VM per host, not all of them

2) This particular path is going to be replaced in a week or two anyway when I migrate from the old fabric switch to the new one, which uses different input ports on our array.   So the "SP A1 and "SP B1" are "going away", to be replaced with "A5" and "B5";  I thought it'd be good though to work from a clean, problem-free starting point.

Can you try to remove powermt.custom file and run "powermt config" again?

Should that not help, would it be possible to try it again, but stop and start PP first?

Many thanks,

Pawel

I'm not sure it's a safe idea to just kill PowerPath on a running host.. Out of the four VMs, this only affects one. I'd already tried deleting the custom_XML file.

I wound up just rebooting them all, to be safe.

Everything looks solid now.

It's disappointing though that in linux (SLES11), a reboot was necessary.