This post is more than 5 years old
1 Rookie
•
82 Posts
1
8550
January 23rd, 2011 23:00
SCSI-3 Persistent Reserve: Enabled & Disabled
hi -would like to check what will differ if i mount below two dev to a host since one dev
is SCSI-3 presistent reserve enabled and other is disabled.
# symdev -sid 25 show 1234 | grep SCSI-3
SCSI-3 Persistent Reserve: Enabled
# symdev -sid 25 show ABCD | grep SCSI-3
SCSI-3 Persistent Reserve: Disabled
No Events found!
Sam Claret
86 Posts
0
January 23rd, 2011 23:00
Hello
You need to add some environmental factors for this question - like the OS type and whether the node is clustered.
All other things being equal in a non clustered host you should have no issue between the 2 devices, however it is not normal to set the SCSI3 persistent reservation flag on a non shared device.
In a clustered node you may well be able to mount both however with certain clusters you will be unable to failover/failback the device with the SCSI3 bit set DISABLED.
I hope this helps?
Sam Claret EMC TSE3
tk_venu
1 Rookie
•
82 Posts
0
January 24th, 2011 02:00
two devs are mapped to a Lunix Guest on a VMware farm.
emcsanadmin
54 Posts
1
January 24th, 2011 18:00
so one lun is set as SCSI-3 Persistent Reserve: Enabled & other one is disabled i think you need to disable the SCSI-3 Persistent Reserve on the other device since the server is not in a cluster mode.
Jurjen_Oskam
56 Posts
0
January 31st, 2011 11:00
Hi,
What will differ is that the device with the SCSI-3 Persistent Reserve flag set to "enabled" will accept and process the SCSI commands PERSISTENT RESERVE IN and PERSISTENT RESERVE OUT, and the device with the flag set to "disabled" will not process those commands.
How this affects your environment depends on whether these commands are ever sent to the devices. If they are not, the differing flags will not make any difference. If they *are* sent, they will fail on the device with the flag set to "disabled", and succeed on the device with the flag set to "enabled". How *this* in turn would affect your environment is hard to tell, since you didn't really specify your environment.
Anyway, I don't think this is the answer you were looking for.
Basically, it comes down to this:
* If you're just testing or experimenting on a non-production system with data of no value, go right ahead and try. Nothing will be physically damaged.
* If this concerns a production system (or a system containing data of any value greater than zero): just Don't Do It. Disable the flag on all devices, and ONLY turn it on when the EMC Interoperability Matrix or other official documentation tells you to (and then only on the correct devices, and make sure that you know which devices are the correct ones and which aren't). This is not meant to scare, but it's just good practice: sure, just setting or clearing flags might work in your current setup, but what will happen with the next version of your OS? Or the next Service Pack? Or the next patch? Doing this is an accident waiting to happen: if the behaviour of the system happens to change without you knowing, you have a problem.
(For what it's worth: enabling the flag is generally only done for some types of clusters.)
Regards,
Jurjen Oskam
dynamox
9 Legend
•
20.4K Posts
0
August 15th, 2013 12:00
Run “prepare” ..it will let you know.
balamurali_kona
2 Posts
0
August 15th, 2013 12:00
HI ,
I am looking to check if to enable or disable SCSI3 for a lun already mapped needs to be unmapped before making any changes .
balamurali_kona
2 Posts
0
August 15th, 2013 12:00
It asked me to unmap in microcode 5876 .But I remember not unmapping previously , not sure about micro code .. wann make sure if this was changed in code upgrade .
dynamox
9 Legend
•
20.4K Posts
1
June 6th, 2017 10:00
set device 0864 attribute=no scsi3_persist_reserv;