GK cannot be opened becase locked or inaccessible

1. Make sure none of your gatekeepers are associated with device groups.
2. Did you have any symcli commands that maybe hung, or didn't complete?
3. Check a symcfg -lockn 13 list. If there's nothing going on and you have a lock in place
you might consider removing the lock.
4. You might need a 3rd gatekeeper depending on how many timefinder/etc operations
are happening simultaneously.
5. Try restarting all symapisrv and/or storapid processes and then try your operation again.

Check out emc136717 and emc107027 as a starting point.

Edit: Also check out emc94039.

Message was edited by:
smw6181

#symcfg list -semaphores.

Lock Proc.
ID Key State Type Wait Lock Full Path Specification
------ ---------- -------- ---- ----- ----------------------------

1867803 0x450ccfea Unlocked GK 0 /dev/rdsk/c36t0d0
524323 0x450ccff5 Unlocked GK 0 /dev/rdsk/c38t0d0
25 0x450cd123 Locked GK 0 /dev/rdsk/c38t0d1
33 0x450cd125 Unlocked GK 0 /dev/rdsk/c38t0d2
98340 0x450d1e90 Unlocked GK 0 /dev/rdsk/c40t0d0
1179685 0x450d1e86 Unlocked GK 0 /dev/rdsk/c43t0d0
23 0x49157442 Unlocked DB 0 /var/symapi/db/symapi_db.bin

:root [/opt/emc/API/symapi/config] powermt display dev=c43t0d0
Bad dev value c43t0d0, or not under Powerpath control.
:root [/opt/emc/API/symapi/config] powermt display dev=c40t0d0
Bad dev value c40t0d0, or not under Powerpath control.

How can i release/cleaup those inaccessibe GK?

use 'ipcrm -s ' with causion.
make sure you got the right number

I also advise killing all stordaemons with 'stordaemon shutdown all -immediate'

why would you not want to associate your gatekeepers with device groups ?. According to "EMC Solutions Enabler Symmetrix Array Management CLI" you do want to associate gatekeepers with a device group so they are used by symcli first instead of regular devices when issuing commands to your Symm.

According to emc136717:

In this particular case, the server did have gatekeepers assigned to it, but all of them had been associated with device groups. The storapid daemon will not select gatekeepers which are in device groups.


I should have been more clear and said, 'make sure at least one GK isn't associated'.

DEBUG enabled on the script . But there are no errors for last three days.

##added on 08142009 for debug
SYMAPI_DEBUG=-1
export SYMAPI_DEBUG
SYMAPI_DEBUG_FILENAME=/eb_rao11_d/applmgr/SYMAPI_DEBUG/SYMAPI_DEBUG.txt
export SYMAPI_DEBUG_FILENAME
## END of debug

can a VCMDB device be a candidate for GK?

Since DMX3 the VCMDB is a GK.

can you point me a doc saying so?

emc128411 Volume Configuration Manager (VCM) device and Enginuity 5771

Edit to add link: http://csgateway.emc.com/primus.asp?id=emc128411

This explains all you need to know about how the VCM works with the latest code releases

Message was edited by:
Dave Yates,
The Benevolent Moderator

You might try Primus 174877 that says:
Fix As a workaround, recreate the ldb directory by using the following steps (assuming S.E. is in the default directory) and retry the command. If root cause is required, contact the Support Center and quote this solution.
1. c:\program files\emc\symcli\storbin\stordaemon shutdown storapid
2. ren c:\program files\emc\symapi\ldb c:\program files\emc\symapi\ldb.bad
3. c:\program files\emc\symcli\storbin\stordaemon start storapid

never knew what step 2 was but did step 1 and 3 and it fixed my issue. I finally had to add gate keepers as you need a minimum of 6. Good luck
MKG

second step is for renaming the Local DataBase . We are not using storapid any more (storapid:client_use = never). I did add one GK more and not seen the problem for a week now. I will HOLD any more changes till we see the problem again. I am keeping a note of this.

emc128411 is not so stright forward to understand .Note that my array is at code 5772.

"Solutions Enabler 6.3 and higher will block the setting of this bit". That does sound like VCMDB is still not used as GK. is that right??

Message was edited by:
SKT

Adding another GK helped the situation. No further occurence for a month

SKT,
it is not advisable to run without tha storapid process running. We only ever recommended implementing this as a workaround.

The daemon does lots of neat stuff, and should not be turned off.

If you open an SR with support we can have a look at your configuration and tell you exactly what you need to do to correct your situation.