Duplicate Disk Access (DA) names observed after performing Dynamic Reconfiguration (DR) steps
book
Article ID: 100023195
calendar_today
Updated On:
Resolution
Problem:
Duplicate DA names were observed in the"vxdisk list" output, after performing the following DR steps.
1) removing devices from vxvm( vxdisk rm) and from OS (rmsf -a device)
2) ioscan
3) vxdiskscandisks new
# vxdisk list
DEVICE TYPE DISK GROUP STATUS
disk4_p2 auto:hpdisk rootdisk01 rootdg online
disk5 auto:cdsdisk - - online
disk6 auto:cdsdisk - - online
disk15 auto - - error<<<<
disk5 auto - - error
disk15 auto:cdsdisk disk15 dg15 online
disk13 auto:cdsdisk disk13 dg13 online
Also paths under two dmpnodesget swapped when more than one device is removed from vxvm and OS as mentionedabove.
c4t0d3 state=enabled type=primary
c8t0d4 state=enabled type=secondary <<<<<<<<<<<<<<
c16t0d4state=enabled type=secondary <<<<<<<<<<<<<<<
c6t0d3 state=enabled type=primary
c10t0d3 state=enabled type=primary
c14t0d3state=enabled type=primary
c14t0d4 state=enabled type=primary<<<<<<<<<<<<<<<<<
c8t0d3 state=enabled type=secondary
The /etc/vx/dmppolicy.info is setup as follows:
dmppolicy.info: START
arraytype
#
arrayname
#
enclosure
#
naming
scheme=osn persistence=yes mode=new
dmppolicy.info: END
A reboot of the machine is required to clear the above state.
Solution:
The issues described above are fixed in an patch via incident 1748068. The patch is built on top of 5.0 GA RP4 and is called 5.0RP4HF1.
Please contact technical support for this fix.
Please note that along with this fix, when performing such DR operations, the dmp_cache_open tunable needs to be set to "on". This is set to"off" by default on HP-UX.
root@mtvrx263 # vxdmpadm get tune all
Tunable Current Value DefaultValue
------------------------------ ------------- -------------
dmp_failed_io_threshold 57600 57600
dmp_retry_count 30 30
dmp_pathswitch_blks_shift 10 10
dmp_queue_depth 40 40
dmp_cache_open off off <<<<<<<<<<<<<<<<<<
dmp_daemon_count 10 10
dmp_scsi_timeout 60 60
dmp_delayq_interval 15 15
dmp_path_age 300 300
dmp_stat_interval 1 1
dmp_health_time 60 60
dmp_probe_idle_lun on on
dmp_log_level 1 1
dmp_retry_timeout 0 0
dmp_fast_recovery off off
dmp_evm_handling off off
dmp_lun_retry_timeout 0 0
To set this to"on", run
# vxdmpadm set tune dmp_cache_open=on
The tunable value will be changed immediately
root@mtvrx263 # vxdmpadm get tune all
Tunable Current Value DefaultValue
------------------------------ ------------- -------------
dmp_failed_io_threshold 57600 57600
dmp_retry_count 30 30
dmp_pathswitch_blks_shift 10 10
dmp_queue_depth 40 40
dmp_cache_open on off <<<<<<<<<<<<<<<<<
dmp_daemon_count 10 10
dmp_scsi_timeout 60 60
dmp_delayq_interval 15 15
dmp_path_age 300 300
dmp_stat_interval 1 1
dmp_health_time 60 60
dmp_probe_idle_lun on on
dmp_log_level 1 1
dmp_retry_timeout 0 0
dmp_fast_recovery off off
dmp_evm_handling off off
dmp_lun_retry_timeout 0 0
Once the DR operations are done, please retain dmp_cache_open to "on".
Issue/Introduction
Duplicate Disk Access (DA) names observed after performing Dynamic Reconfiguration (DR) steps
Was this article helpful?
thumb_up
Yes
thumb_down
No