Asset ID: |
1-72-1357486.1 |
Update Date: | 2012-06-21 |
Keywords: | |
Solution Type
Problem Resolution Sure
Solution
1357486.1
:
Attempt to unmap LUNs met with Hitachi Storage Navigator Error Number: 2082; Error Message: Processing is disabled because the host is mounted or the device is reserved.
Related Items |
- Sun Storage 9985 System
- Sun Storage 9990 System
- Sun Storage 9990V System
- Sun Storage 9970 System
- Sun Storage 9985V System
- Sun Storage 9980 System
|
Related Categories |
- PLA-Support>Sun Systems>DISK>HighEnd Arrays>SN-DK: HDS_99xx
- .Old GCS Categories>Sun Microsystems>Storage - Disk>Datacenter Disk
|
In this Document
Created from <SR 3-4344544211>
Applies to:
Sun Storage 9990 System - Version Not Applicable and later
Sun Storage 9985 System - Version Not Applicable and later
Sun Storage 9990V System - Version Not Applicable and later
Sun Storage 9985V System - Version Not Applicable and later
Sun Storage 9980 System - Version Not Applicable and later
Information in this document applies to any platform.
Symptoms
An attempt to unmap a LUN or LUNs from a host group using Hitachi Storage Navigator or Device Manager is met with the following error:
Error Number: 2082
Error Message: Processing is disabled because the host is mounted or the device is reserved.
Recommended Action: The host is mounted or the device is reserved. Please unmount the host and cancel reservation of the device.
Severity Level: E
Failure Level: 6
Cause
The attempt to unmap the LUNs using Hitachi Storage Navigator or Device Manager cannot be completed because the fabric devices are still configured on the Solaris host.
Solution
1. Stop all device activity to each fabric device to be unmapped.
2. Unmount any file systems associated with the fabric devices to be unmapped.
3. Remove any SCSI reservations associated with the fabric devices to be unmapped.
4. If the fabric devices to be unmapped are under control of a volume manager control, reference the documentation of the associated volume manager for maintaining the fabric device.
5. Unconfigure the fabric devices associated with Sun StorageTek Traffic Manager Multipathing-Enabled Devices. Reference the Chapter "Unconfiguring Fabric Devices" in Appendix A of the Solaris SAN Configuration and Multipathing Guide as well as the cfgadm_fp man page.
6. Reattempt to unmap the LUNs in Hitachi Storage Navigator or Device Manager.
Internal Only
There are documents that explains how to solve this issue from the SVP of the Storage array.
This should be done only by Oracle engineers trained on HDS products.
1) For (SE9910, SE9960, SE9970 and SE9980)
--------------------------------------
Information Regarding the Force Reset Function for Reserves on the 9900V
ID: hds59318
Question: Reallocate or remove LUNs with reserves from hosts that were improperly removed
Environment: R450
9900V
Host Reserve
Force Reset
Debug Mode
Reallocate
LUN
Resolution:
There is no password release/reset function for clearing Host Reserves on any Raid models prior to the R500 (USP). If you have orphaned Host reserves that need to be reset on 9980 LUNs you have two options:
1. Using Debug mode issue force resets to the port that the LUNs in question are mapped to.
Note: This is disruptive to the port and can potentially reset valid reserves so you have to take every precaution when doing this. It is possible and recommended that if you can not take all of the hosts offline to the port(s) where you need to issue the force reset, which is required to guarantee data integrity, then you can map the LUNs in question to an unused port and issue the force reset on that port.
2. POR the subsystem - obviously disruptive as well.
There are no other options to clear orphaned reserves on Raid model subsystems prior to the USP.
--------------------------------------
Procedure to issue force resets using Debug Mode, only for R400 and R450 (SE9910, SE9960, SE9970 and SE9980):
Unfortunately we do not have the password-based facility for selectively removing reservations on older arrays (R450 and older). See more for action plan to release reservations. Note that it is possible that the following procedure will not work on persistent reservations. This depends on the options (bit flags set by the driver) that were selected when the reservations were first established. Persistent reservations can be configured to survive power cycles and all types of resets. If this procedure does not work then the reservations will have to be removed by an attached host using appropriate SCSI-3 commands. This does not have to be done by the host that originally established the reservations. 1) Map the LUNs from which reservations need to be forcibly removed to an unused port. The customer should be able to do this using storage navigator. The following procedure should be run on the port selected by the customer for a forcible reset.
WARNING: ALL Hosts that have access to the port to be reset MUST be down, preferably powered off. (This is to insure that the port does not detect ANY IO activity of any kind.)
The following is the detailed procedure for doing a FORCE RESET to clear any reserves for the selected LUNs. To prepare for this, the host(s) should be down that is connected to the interfaces that are to be reset. If a Force Reset is issued to an interface that is operating to a host(s), the reset will cause an interface reset and it will disrupt the host(s). It is preferable to have the host(s) powered down to do the Force Reset and ensure the proper completion.
* Start from view mode on SVP screen.
* Enter CTRL/SHIFT and D keys together.
* Enter password RAID-DEBUG
* This will change mode from View to Debug Mode and allows the DEBUG function selection.
* Select DEBUG and the DEBUG window will display the selection options. Select FORCE RESET.
* This displays a FORCE RESET window with a selection box for each SCSI port/channel. This reset only works on SCSI and Fibre interfaces and is not valid for ESCON. Use the mouse to enter a check mark in the boxes selecting the interfaces you wish to reset.
* Select OK to issue the reset.
* A message will display FORCE RESET FINISHED and will have an OK entry that you will do.
* Select FORCE RESET again from the Debug selection window.
* The FORCE RESET window will display again with the check mark gone if the reset was actually successful. If the check mark is still displayed in the box, that interface was not reset and the procedure will need to be repeated.
(NOTE: the check mark MAY still be displayed in the box. Check the SSB log for presence of EC 1693 logged by the MP controlling the port you selected. This SSB indicates the reserve was released. TRC has been told that it should take a maximum of about 15 minutes for the SSB to appear. Close the Debug Mode window, and return the SVP to View Mode. Then attempt to remove the LUN from the SCSI configuration using the usual methods at the SVP or Remote Console.)
* When all the required interfaces have been reset, exit Debug Mode window.
* Select the MODE button to return to View mode.
NOTE: This procedure may need to be run more then once to clear the reserves. If the reserve flag is still set after following this procedure, please repeat the steps again until the reserve is cleared.
2) For SE9985, SE9990, ST9990V and ST9985V
First check if there are scsi reservations on the LUNs :
Determine LUN Status on TagmaStore USP
ID: hds55742
Date Created: 11/28/2006
Last Modified: 06/17/2009
Status: Approved
Validated: By Review
Question: |
How to find if the LUN has reserve |
|
Lun Status |
|
Lun Reserve |
|
Display Reserve |
|
Display Reserves |
|
Display Lun Status |
|
Check Lun Status |
Environment: |
Tagmastore USP |
|
Persistent Reserve |
|
Reserve |
|
Status |
Resolution: |
If you select LUN Status from View in the LUN Manager panel on the SVP, the LUN status will be displayed in the LUN list in the panel.
The following values can appear in the Status column.
Status
|
Explanation
|
Normal
|
Normal device
|
BLK
|
It is not ready due to blockade.
|
OPR
|
It is reserved by the normal OpenReserve command.
|
KEY
|
Persistent Group Reserve key is set.
|
PGR
|
It is reserved by the Persistent Group Reserve command.
|
MFR
|
It is reserved by Mainframe.
|
H35R
|
It is reserved from the H3500 server.
|
ACA
|
It is in the ACA ACTIVE status.
|
Note: ACA means Auto Contingent Allegience, used with AIX. Details on ACA.
|
Once confirmed there are scsi reservations, then follow
LUN reserved status needs to be Force Reset (RAID500)
ID: hds48343
Date Created: 06/17/2005
Last Modified: 02/09/2006
Status: Approved
Validated: In Lab
Problem: |
LUN reserved
|
|
LUN status is reserved
|
Environment: |
Force reset
|
|
TagmaStore USP
|
|
TagmaStore NSC55
|
|
Need to reset LUN
|
Resolution: |
LUN-level force reset, as opposed to Port-level force reset, is supported on the RAID500 model.
Open a Clarify case for tracking and reporting.
Contact TRC for procedure and one-time password to reset LUN reserved status.
Refer TRC to this solution for access to further instructions.
|
Procedure to do a LUN-level force reset on the SVP of the storage array:
- Change to Modify Mode.
- Install -> Change Configuration -> LUN Configuration
- View -> LUN Status (Check Enabled)
- Select desired port
- Change Display from Group to LUN
- Select desired host group
- Select desired LUN or LUNs
- LUN -> Force Reset (desde Lun Management --> Lun status)
- Enter time sensitive password into PAS3360W dialog window and
select OK.
Other internal only resource are SAN chat room "disk-highend"
As well as, [email protected] Beehive Group alias
Attachments
This solution has no attachment