Method of allowing a volume manager to release and regain reserve

• US 6,526,492 B1
• Filed: 06/30/1998
• Issued: 02/25/2003
• Est. Priority Date: 06/30/1998
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: McBrearty, Gerald Francis, Shieh, Johnny Meng-Han
• Primary Examiner(s): Lane, Jack A.
• Application Number:
  US09/107,104
• Time in Patent Office: 1,701 Days
• Field of Search: 711/4, 711/111, 711/170
• US Class Current: 711/170
• CPC Class Codes: G06F 2003/0697 device management, e.g. handlers, drivers, I/O schedulers
• CPC Class Codes: G06F 3/0601 Interfaces specially adapted for storage systems

Abstract

In a multiple machine data processing system, a volume group may be accessed by other than the original owner of the volume group and maintain integrity of the volume group. A logical volume manager on the primary machine holds all incoming I/O requests to the logical volumes in the volume group and waits for all the I/O requests already sent down to the disks in the volume group. Once all the outstanding I/O requests to the disk have been completed, the disks in the volume group are closed and reopened without reserve. The logical volume manager on a secondary machine opens the disks in the volume group without taking a reserve to allow meta-data to be refreshed on the secondary machine. When the secondary machine is finished the disks are closed. The logical volume manager on the primary machine holds all incoming I/O requests to the logical volumes in the volume group and waits for all the I/O requests already sent down to the disks in the volume group. Once all the outstanding I/O requests to the disks have been completed, the disks in the volume group are closed and reopened with reserve. The application on the primary machine does not see any of the operations and is unaffected.

Claims

1. A method for managing a storage device utilized by multiple machines in a non-concurrent mode, comprising:

controlling a reserve for a first machine in a volume group by;

releasing the reserve for the first machine in the volume group;

accessing the disks in the volume group by a second machine;

re-establishing the reserve for the first machine in the volume group; and

suspending input/output requests for the first machine while releasing the reserve, accessing the disks and re-establishing the reserve.

2. The method of claim 1, wherein controlling the reserve on a volume group for a first machine, further comprises:

holding new input/output requests for the volume group on the first machine;

finishing current input/output requests for the volume group on the first machine;

closing the storage device and

releasing the reserve on the volume group for the first machine;

re-opening the storage device without the reserve on the first machine; and

releasing retained input/output requests for the first machine.

3. The method of claim 1, wherein accessing the disks in the volume group by a second machine, further comprises:

opening a volume group on the second machine without the reserve;

determining the changes made in the volume group by the first machine;

updating meta-data on the second machine; and

closing the volume group on the second machine.

4. The method of claim 1, wherein reopening the reserve to the first machine, further comprises:

holding new input/output requests for the volume group on the first machine;

finishing current input/output requests for the volume group on the first machine;

closing the storage device and

re-establishing the reserve;

re-opening the storage device and re-establishing the reserve on the first machine; and

releasing retained input/output requests for the first machine.

5. A data processing system for managing a storage device utilized by multiple machines in a non-concurrent mode, comprising:

a logical volume manager for controlling a reserve on disks in a volume group for a first machine, further comprising;

release means for releasing the reserve for the first machines in the volume group;

access means for accessing the disks in the volume group by a second machine; and

re-establishment means for re-establishing the reserve for the first machine in the volume group; and

a buffer for suspending input/output requests for the first machine while releasing the reserve, accessing the disks and re-establishing the reserve.

6. The data processing system of claim 5, wherein the buffer for holding input/output requests for the first machine, further comprises

conclusion means for finishing current input/output requests for the first machine.

7. The data processing system of claim 5, wherein the logical volume manager for controlling a reserve on a volume group for a first machine, further comprises:

an opening means for opening a volume group on the second machine;

an update means for updating the local meta-data on the second machine with changes to the volume group on the first machine; and

closure means for closing the volume group on the second machine.

8. The data processing system claim 7, further comprises:

suspension means for suspending new input/output requests for the volume group on the first machine;

finishing means for finishing current input/output requests for the volume group on the first machine;

means for re-establishing the reserve on the first machine;

means for closing the storage device and

means for re-opening the reserve on the storage device to the first machine; and

release means for releasing retained input/output requests for the first machine.