power path and VG alternate path

[skt_skt]

Any one have a working combination of Clarrion model (cx700) , HP-UX 10.20 and a power path s/w .

We had observed a tree passing issue on this setup and power path(dont know the exact version) module was removed (one year back)as it is reported unsupported on HP-UX 10.20.

Now we have two HBA in the system and 10 LUN is asigned to HBA1 remaining 10 LUN to HBA2 for load balancing.Also each of the 10 LUN is zoned to one HBA only so that only path is visible for each LUN.

I would like to keep all the LUNs zoned to both HBA so that the server is able to see two paths which can be added in VG as an alternate path.Currently the VG holds only one path of the each LUN.

The question is there a PP version which can work here, even though not supported? Or without PP can the VG or system handle the a tress pass in case of a SP failure or a HBA failure.

[Geoff Wild]

When running Powerpath - you do NOT use alternate links.

Powerpath will look after that for you.

Rgds...Geoff

[skt_skt]

"We had observed a tress passing issue on this setup and power path(dont know the exact version) module was removed (one year back)as it is reported unsupported on HP-UX 10.20"

I dont have a PP installed now. How the system/VG will do the switching to the alternate path w/o management software, in case we dont have any supported PP.

[Tim Nelson · Accepted solution]

PVLinks or alternate pathing is built into LVM.

If no multipathing software is installed then simply add the other path to your volume group(s). LVM will recognize that it is the same disk just different path and label them as alternate paths.

I there is a failure to the primary path LVM will switch to the alternate path. Once the primary path is fixed I do not believe LVM will switch it back. You will have to do this manually.

You are running HPUX 10.20 so many things may or may not work perfectly for you due to the age of the OS and lack of updates and/or patches.

[psreedhar]

Hi,
LVM PVlinks concept is not exact replacement for PP.

PP does load balencing also, but LVM doesn't.

[skt_skt]

we are planning the load balancing in the VG level.Half of the disks in the VG will use one path and other half will use another path. What ever i posted below we will be adding the alternate path with vgextend.

--- Physical volumes ---
PV Name /dev/dsk/c46t0d3
PV Status available
Total PE 2157
Free PE 0
Autoswitch On

PV Name /dev/dsk/c46t0d7
PV Status available
Total PE 2157
Free PE 0
Autoswitch On

PV Name /dev/dsk/c46t1d3
PV Status available
Total PE 2157
Free PE 0
Autoswitch On

PV Name /dev/dsk/c47t0d5
PV Status available
Total PE 2157
Free PE 0
Autoswitch On

PV Name /dev/dsk/c47t1d1
PV Status available
Total PE 2157
Free PE 0
Autoswitch On

PV Name /dev/dsk/c47t1d5
PV Status available
Total PE 2157
Free PE 942
Autoswitch On

[Michael Steele_2]

Load balancing in round robin algorithm is the default allocation in LVM. See distributed allocation policy for 'lvcreate'.

Set the distributed allocation policy. distributed can have one of the following values:

y Turn on distributed allocation.

n Turn off distributed allocation. This is the default.



When the distributed allocation policy is turned on, only one free extent is allocated from the first available physical volume. The next free extent is allocated from the next available physical volume. Allocation of free extents proceeds in round-robin order on the list of available physical volumes.

When the distributed allocation policy is turned off, all available free extents are allocated from each available physical volume before proceeding to the next available physical volume. This is the default.

[skt_skt]

now both the HBA are zoned to see all the LUNs. then ran the ioscan and insf.

But i am still not able to access the secondory path. what you feel lokking at the o/p?

Here is the new inq o/p

r#oot [/root] /usr/local/bin/inq
Inquiry utility, Version V7.2-154 (Rev 4.0) (SIL Version V4.2-154)
Copyright (C) by EMC Corporation, all rights reserved.
For help type inq -h.

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

-------------------------------------------------------------------------
DEVICE :VEND :PROD :REV :SER NUM :CAP(kb)
-------------------------------------------------------------------------
/dev/rdsk/c0t5d0 :SEAGATE :ST318436LC :HP05 :3BM0VMJ8 :17783240
/dev/rdsk/c0t6d0 :SEAGATE :ST318436LC :HP05 :3BM0W30M :17783240
/dev/rdsk/c1t2d0 :HP :DVD-ROM 304 :2.00 : 1999/10 : 190380
/dev/rdsk/c46t0d0 :DGC :CX700WDUNB :HP03 :Ch2 CONT : 0
/dev/rdsk/c46t0d1 :DGC :CX700WDR5 :HP03 :Ch2 CONT :26516160
/dev/rdsk/c46t0d2 :DGC :CX700WDR5 :HP03 :Ch2 CONT : 0
/dev/rdsk/c46t0d3 :DGC :CX700WDR5 :HP03 :Ch2 CONT :8838720
/dev/rdsk/c46t0d4 :DGC :CX700WDR5 :HP03 :Ch2 CONT : 0
/dev/rdsk/c46t0d5 :DGC :CX700WDR5 :HP03 :Ch2 CONT : 0
/dev/rdsk/c46t0d6 :DGC :CX700WDR5 :HP03 :Ch2 CONT :8838720
/dev/rdsk/c46t0d7 :DGC :CX700WDR5 :HP03 :Ch2 CONT :8838720
/dev/rdsk/c46t1d0 :DGC :CX700WDR5 :HP03 :Ch2 CONT : 0
/dev/rdsk/c46t1d1 :DGC :CX700WDR5 :HP03 :Ch2 CONT : 0
/dev/rdsk/c46t1d2 :DGC :CX700WDR5 :HP03 :Ch2 CONT :8838720
/dev/rdsk/c46t1d3 :DGC :CX700WDR5 :HP03 :Ch2 CONT :8838720
/dev/rdsk/c46t1d4 :DGC :CX700WDR5 :HP03 :Ch2 CONT : 0
/dev/rdsk/c46t1d5 :DGC :CX700WDR5 :HP03 :Ch2 CONT : 0
/dev/rdsk/c46t1d6 :DGC :CX700WDR5 :HP03 :Ch2 CONT :8838720
/dev/rdsk/c46t1d7 :DGC :CX700WDR5 :HP03 :Ch2 CONT :8838720
/dev/rdsk/c46t2d0 :DGC :CX700WDR5 :HP03 :Ch2 CONT : 0
/dev/rdsk/c46t2d1 :DGC :CX700WDR5 :HP03 :Ch2 CONT : 0
/dev/rdsk/c46t2d2 :DGC :CX700WDR5 :HP03 :Ch2 CONT :8838720
/dev/rdsk/c46t2d3 :DGC :CX700WDR5 :HP03 :Ch2 CONT :8838720
/dev/rdsk/c46t2d4 :DGC :CX700WDR5 :HP03 :Ch2 CONT : 0
/dev/rdsk/c46t2d5 :DGC :CX700WDR5 :HP03 :Ch2 CONT : 0
/dev/rdsk/c46t2d6 :DGC :CX700WDR5 :HP03 :Ch2 CONT :8838720
/dev/rdsk/c46t2d7 :DGC :CX700WDR5 :HP03 :Ch2 CONT :8838720
/dev/rdsk/c46t3d0 :DGC :CX700WDR5 :HP03 :Ch2 CONT : 0
/dev/rdsk/c46t3d1 :DGC :CX700WDR5 :HP03 :Ch2 CONT : 0
/dev/rdsk/c46t3d2 :DGC :CX700WDR5 :HP03 :Ch2 CONT :8838720
/dev/rdsk/c46t3d3 :DGC :CX700WDR5 :HP03 :Ch2 CONT :8838720
/dev/rdsk/c46t3d4 :DGC :CX700WDR5 :HP03 :Ch2 CONT : 0
/dev/rdsk/c46t3d5 :DGC :CX700WDR5 :HP03 :Ch2 CONT : 0
/dev/rdsk/c46t3d6 :DGC :CX700WDR5 :HP03 :Ch2 CONT :8838720
/dev/rdsk/c46t3d7 :DGC :CX700WDR5 :HP03 :Ch2 CONT :8838720
/dev/rdsk/c46t4d0 :DGC :CX700WDR5 :HP03 :Ch2 CONT : 0
/dev/rdsk/c46t4d1 :DGC :CX700WDR5 :HP03 :Ch2 CONT : 0
/dev/rdsk/c46t4d2 :DGC :CX700WDR5 :HP03 :Ch2 CONT :8838720
/dev/rdsk/c46t4d3 :DGC :CX700WDR5 :HP03 :Ch2 CONT :8838720
/dev/rdsk/c46t4d4 :DGC :CX700WDR5 :HP03 :Ch2 CONT : 0
/dev/rdsk/c46t4d5 :DGC :CX700WDR5 :HP03 :Ch2 CONT : 0
/dev/rdsk/c46t4d6 :DGC :CX700WDR5 :HP03 :Ch2 CONT :8838720
/dev/rdsk/c46t4d7 :DGC :CX700WDR5 :HP03 :Ch2 CONT :8838720
/dev/rdsk/c46t5d0 :DGC :CX700WDR5 :HP03 :Ch2 CONT : 0
/dev/rdsk/c46t5d1 :DGC :CX700WDR5 :HP03 :Ch2 CONT : 0
/dev/rdsk/c46t5d2 :DGC :CX700WDR5 :HP03 :Ch2 CONT :8838720
/dev/rdsk/c46t5d3 :DGC :CX700WDR5 :HP03 :Ch2 CONT : 0
/dev/rdsk/c46t5d4 :DGC :CX700WDR5 :HP03 :Ch2 CONT :8838720
/dev/rdsk/c47t0d0 :DGC :CX700WDUNB :HP03 :Ch2 CONT : 0
/dev/rdsk/c47t0d1 :DGC :CX700WDR5 :HP03 :Ch2 CONT : 0
/dev/rdsk/c47t0d2 :DGC :CX700WDR5 :HP03 :Ch2 CONT :17677440
/dev/rdsk/c47t0d3 :DGC :CX700WDR5 :HP03 :Ch2 CONT : 0
/dev/rdsk/c47t0d4 :DGC :CX700WDR5 :HP03 :Ch2 CONT :8838720
/dev/rdsk/c47t0d5 :DGC :CX700WDR5 :HP03 :Ch2 CONT :8838720
/dev/rdsk/c47t0d6 :DGC :CX700WDR5 :HP03 :Ch2 CONT : 0
/dev/rdsk/c47t0d7 :DGC :CX700WDR5 :HP03 :Ch2 CONT : 0
/dev/rdsk/c47t1d0 :DGC :CX700WDR5 :HP03 :Ch2 CONT :8838720
/dev/rdsk/c47t1d1 :DGC :CX700WDR5 :HP03 :Ch2 CONT :8838720
/dev/rdsk/c47t1d2 :DGC :CX700WDR5 :HP03 :Ch2 CONT : 0
/dev/rdsk/c47t1d3 :DGC :CX700WDR5 :HP03 :Ch2 CONT : 0
/dev/rdsk/c47t1d4 :DGC :CX700WDR5 :HP03 :Ch2 CONT :8838720
/dev/rdsk/c47t1d5 :DGC :CX700WDR5 :HP03 :Ch2 CONT :8838720
/dev/rdsk/c47t1d6 :DGC :CX700WDR5 :HP03 :Ch2 CONT : 0
/dev/rdsk/c47t1d7 :DGC :CX700WDR5 :HP03 :Ch2 CONT : 0
/dev/rdsk/c47t2d0 :DGC :CX700WDR5 :HP03 :Ch2 CONT :8838720
/dev/rdsk/c47t2d1 :DGC :CX700WDR5 :HP03 :Ch2 CONT :8838720
/dev/rdsk/c47t2d2 :DGC :CX700WDR5 :HP03 :Ch2 CONT : 0
/dev/rdsk/c47t2d3 :DGC :CX700WDR5 :HP03 :Ch2 CONT : 0
/dev/rdsk/c47t2d4 :DGC :CX700WDR5 :HP03 :Ch2 CONT :8838720
/dev/rdsk/c47t2d5 :DGC :CX700WDR5 :HP03 :Ch2 CONT :8838720
/dev/rdsk/c47t2d6 :DGC :CX700WDR5 :HP03 :Ch2 CONT : 0
/dev/rdsk/c47t2d7 :DGC :CX700WDR5 :HP03 :Ch2 CONT : 0
/dev/rdsk/c47t3d0 :DGC :CX700WDR5 :HP03 :Ch2 CONT :8838720
/dev/rdsk/c47t3d1 :DGC :CX700WDR5 :HP03 :Ch2 CONT :8838720
/dev/rdsk/c47t3d2 :DGC :CX700WDR5 :HP03 :Ch2 CONT : 0
/dev/rdsk/c47t3d3 :DGC :CX700WDR5 :HP03 :Ch2 CONT : 0
/dev/rdsk/c47t3d4 :DGC :CX700WDR5 :HP03 :Ch2 CONT :8838720
/dev/rdsk/c47t3d5 :DGC :CX700WDR5 :HP03 :Ch2 CONT :8838720
/dev/rdsk/c47t3d6 :DGC :CX700WDR5 :HP03 :Ch2 CONT : 0
/dev/rdsk/c47t3d7 :DGC :CX700WDR5 :HP03 :Ch2 CONT : 0
/dev/rdsk/c47t4d0 :DGC :CX700WDR5 :HP03 :Ch2 CONT :8838720
/dev/rdsk/c47t4d1 :DGC :CX700WDR5 :HP03 :Ch2 CONT :8838720
/dev/rdsk/c47t4d2 :DGC :CX700WDR5 :HP03 :Ch2 CONT : 0
/dev/rdsk/c47t4d3 :DGC :CX700WDR5 :HP03 :Ch2 CONT : 0
/dev/rdsk/c47t4d4 :DGC :CX700WDR5 :HP03 :Ch2 CONT :8838720
/dev/rdsk/c47t4d5 :DGC :CX700WDR5 :HP03 :Ch2 CONT :8838720
/dev/rdsk/c47t4d6 :DGC :CX700WDR5 :HP03 :Ch2 CONT : 0
/dev/rdsk/c47t4d7 :DGC :CX700WDR5 :HP03 :Ch2 CONT : 0
/dev/rdsk/c47t5d0 :DGC :CX700WDR5 :HP03 :Ch2 CONT :8838720
/dev/rdsk/c47t5d1 :DGC :CX700WDR5 :HP03 :Ch2 CONT :8838720
/dev/rdsk/c47t5d2 :DGC :CX700WDR5 :HP03 :Ch2 CONT : 0
/dev/rdsk/c47t5d3 :DGC :CX700WDR5 :HP03 :Ch2 CONT :8838720
/dev/rdsk/c47t5d4 :DGC :CX700WDR5 :HP03 :Ch2 CONT : 0


am worried if the DVS_UNOPENED is causing the problem. let me know your openion.

:root [/root] /opt/fc/bin/fcmsutil /dev/td0 get remote all

Target N_Port_id is = 0x615e13
Target state = DVS_READY
Target Port World Wide Name = 0x5006016030603525
Target Node World Wide Name = 0x50060160b0603525

Common Service parameters (all values shown in hex):
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Version_Hi: 20 Version_Low: 9
BB_Credit: 8 Ciro: 1
Alt_BBcred: 0 Rxsz: 800
Conseq: ff RO_Bitmap: fe
Resolution: 0 EDTOV(ms): 7d0

Vendor version (in hex) 0 0 0 0
-------------------------------------------------------------


Target N_Port_id is = 0x616213
Target state = DVS_UNOPENED
Target Port World Wide Name = 0x5006016830603525
Target Node World Wide Name = 0x50060160b0603525

Common Service parameters (all values shown in hex):
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Version_Hi: 20 Version_Low: 9
BB_Credit: 8 Ciro: 1
Alt_BBcred: 0 Rxsz: 800
Conseq: ff RO_Bitmap: fe
Resolution: 0 EDTOV(ms): 7d0

Vendor version (in hex) 0 0 0 0
-------------------------------------------------------------

:root [/root] /opt/fc/bin/fcmsutil /dev/td1 get remote all

Target N_Port_id is = 0x625e13
Target state = DVS_UNOPENED
Target Port World Wide Name = 0x5006016130603525
Target Node World Wide Name = 0x50060160b0603525

Common Service parameters (all values shown in hex):
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Version_Hi: 20 Version_Low: 9
BB_Credit: 8 Ciro: 1
Alt_BBcred: 0 Rxsz: 800
Conseq: ff RO_Bitmap: fe
Resolution: 0 EDTOV(ms): 7d0

Vendor version (in hex) 0 0 0 0
-------------------------------------------------------------


Target N_Port_id is = 0x626213
Target state = DVS_READY
Target Port World Wide Name = 0x5006016930603525
Target Node World Wide Name = 0x50060160b0603525

Common Service parameters (all values shown in hex):
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Version_Hi: 20 Version_Low: 9
BB_Credit: 8 Ciro: 1
Alt_BBcred: 0 Rxsz: 800
Conseq: ff RO_Bitmap: fe
Resolution: 0 EDTOV(ms): 7d0

Vendor version (in hex) 0 0 0 0
-------------------------------------------------------------

the switch is McData ED64m.

Let me know your suggetions

[Michael Steele_2]

DVS_UNOPENED is proper state.

[skt_skt]

then what is the difference with DVS_READY?

[Michael Steele_2]

I think the only thing you can do at this point is to enable debugging messages for syslog.log. Ordinarily the HBA initiator with the iscsiutil command can be used, but you're only running 10.20.

Verify log entries made with/in the EMS logs. /etc/opt/resmon/*.log.

Regarding the difference between DVS_UNOPENED and DVS_READY, I'm not able to find an exact definition. I can only tell you that they are proper messages collect by the 'fcmsutil get remote all' command, which goes through the initiator to gather up this data. See 'ioscan -fnC ctl'.

Finally, can you site your references that state 10.20 will work with your HBA's, local and remote? Much of the 10.20 command set just isn't sophisticated enough, that's why your seeing so many responses about load balancing with round robin, etc.

[skt_skt]

here are the logs after the last reboot.and there are lot of duplicate tntries


-------------------Start Event--------------------
Event 391 occurred at Wed Mar 5 13:32:41.947816 2008
Process ID: 3081 (/etc/opt/resmon/lbin/p_client) Log Level: Audit
Request for resource "/storage/events/disks/default/8_4_1_0.97.94.19.0.0.0" by client "HP_HWE_SCP"
has been rejected due to RM_NOT_READY. Request will be re-issued later.


-------------------Start Event--------------------
Event 391 occurred at Wed Mar 5 13:32:41.998233 2008
Process ID: 3081 (/etc/opt/resmon/lbin/p_client) Log Level: Audit
Request for resource "/adapters/events/FC_adapter/8_4_1_0.97" by client "HP_HWE_SCP"
has been rejected due to RM_NOT_READY. Request will be re-issued later.
-------------------End Event----------------------

[Michael Steele_2]

Well it does seem to be application related according to event 391's description:

391: ZTMF-EVT-RM-SetBrSlotArrayFail
The maximum resource manager branch slot array size was exceeded while
reinstating in-doubt transaction branches. A new resource manager branch slot array
size is generated.
Additional Unconditional Tokens
ZTMF-TKN-EVTHDR-PROCESSTYPE
identifies the TMF process that generated the event. The value of this token for
this event is ZTMF-VAL-PROC-TMP.
ZTMF-TKN-RM-SLOTARRAYSIZE
is the subject token of the event message, and it specifies the new resource
manager branch slot array size.
Effect
The resource manager branch slot array size is changed to the specified value.
Recommended Action
Informative message only; no corrective action is needed.
Additional Unconditional Tokens
ZTMF-TKN-EVTHDR-PROCESSTYPE token-type ZSPI-TYP-ENUM.
ZTMF-TKN-RM-SLOTARRAYSIZE token-type ZSPI-TYP-INT.
Event-Message Text
The maximum RM branch slot array size has been exceeded due
to reinstating in-doubt transaction branches. The new
maximum size is ZTMF-TKN-RM-SLOTARRAYSIZE.


Instead of using 'ioscan' from the host have you check the clarion disk array for any failed drives?

[skt_skt]

"Informative message only; no corrective action is needed".. so this is not causing any problem

[skt_skt]

DVS_READY means that the storage device allow initiator logins on the port; DVS_UNOPENED may appear in cases where the WWN of the hba is not registered on a Clariion Array or when a DMX storage port (VCM Enabled) is accesible but no device was masked for that particular port or even in cases where a storage port was moved in SAN from one port-id to another so the host does not authenticate it anymore.
Generally, it means that something is happening with that device but this may only be interpreted on an active device: if the host does not try to acces luns on that specific storage it may be possible that the state of the device to be unopened.
I guess that this state means that the host did not tried to open (login in fabric terms) the devi ce in the last period of time (AFAIK the hp-ux logins to the device each time it opens it) or tried and failed.