ACSLS: CAP Number 1 Is Assigned To A Partition Which Does Not Exist Any More

Asset ID:	1-72-2247222.1
Update Date:	2018-01-16
Keywords:

Solution Type Problem Resolution Sure

Solution 2247222.1 : ACSLS: CAP Number 1 Is Assigned To A Partition Which Does Not Exist Any More

Applies to:

Sun StorageTek SL8500 Modular Library System - Version All Versions and later
Information in this document applies to any platform.

Symptoms

A partitioned SL8500 library, running on FRS_8.51, attached to ACSLS 7.3.1, had the partitioning removed with the HLI2 partition still holding a reservation on CAP B.

SLConsole CAP Table
Address Magazines Present Device State Locked Enabled Cap Status Reserved Reserver Cap Position CAP Access Door Position
1.2.23.1.0 3 Ready true true closed true hli2 N/A N/A
1.2.23.2.0 3 Ready true true closed false none N/A N/A

Query CAP on ACSLS displays State = offline, Status = Audit for CAP B . Audit is an unexpected Status since CAP B is offline

ACSSA> q cap all
2017-03-02 11:04:26 CAP Status
Identifier Priority Size State Mode Status
0, 1, 0 0 39 offline manual Audit
0, 1, 1 0 39 offline manual available
ACSSA>

The pending reservation on CAP B by shadow HLI2 partition never caused more trouble to customer until they upgraded to ACSLS 8.4

Problems started after they moved to the new ACSLS 8.4 server. ACSLS could not start because CAP 0,1,1 was reserved to partition HLI2.

2017-02-28 14:35:09 VARY[37]:
2138 E cl_is_cap_resv.c Unknown 109
cl_is_cap_resv: CAP 0,1,1 is already reserved in partition 2 by host 1.

In addition it was noted that the descriptors for the CAP but also for the Drives and Cells in/hbc/configuration/PartitionConfigurationV3.xml are not valid.

Version 1

-2
3
false
hli
false
3
left6
CAP,1|31,2|4,7|55,1|2,*
0
DRIVE,1|31,1|4,-2|2,1,*
50
CELL,1|31,1|4,-55|55,1|2,*
200
NonPartitionedLibrary
0

Cause

The partitioning was removed with the partition still holding a reservation on a CAP.

Solution

The reservation is stored in the database. Clearing a reservation on CAP requires to clear the database in the library.

Clearing the database is restricted to service user, hence requires an engineer to visit. Please open a MOS Service Request (SR) to schedule the onsite service.

Clearing the database causes the library to reboot. This is a disruptive command so you will need to schedule a downtime.

Then a physical audit of all available slots will need to be run, to re-populate the cartridge database. A physical audit on a one expansion library should not take more than 30 minutes.

Please take note that you have to wait for the physical audit to complete and the cartridge database to be re-populated, for the library to be ready for online operations.

A one hour maintenance window should be enough to complete all steps:

On ACSLS, vary offline the ACS and port
Clear the database -> Reboot and Physical audit
After the reboot is completed check the CAP reservation was cleared.

The descriptor for the CAP is not correct in the Partition.configuration file. The file needs to be re-converted with valid entries, for the library to return a good descriptor for the CAP to ACSLS.

After the Audit is completed run the "Apply Default Capacity"
Verify:
  SLC -> Library -> General Properties will also show Active Cell count reflects the expected capacity of the library.
  SLC -> Library -> General Status screen shows that the Access State is online.
  If not, "Bring Online" via the "-select-transition" to bring the library online.

When all done and green on the SLConsole, vary Port and ACS online on ACSLS. Do a query on resources: lsm, cap and drive to check all is online.

I would recommend you run an ACSLS audit on the ACS to ensure library and ACSLS databases are in sync before you go live.

Then you should be good to go move to the ACSLS 8.4 server.

References

<NOTE:2197923.1> - SL8500 - Amendment to 8.51 Upgrade Procedures

Attachments

This solution has no attachment