Previous | Contents | Index |
Use the following syntax of the DCL command MOUNT to mount a shared disk for single or clustered systems:
$ MOUNT/SYSTEM/NOCACHE MDB150 FASTRAMDISK |
The default for a normal hard disk when using the MOUNT command is to the CACHE qualifier. Therefore, you need to specify the /NOCACHE qualifier when using a RAM disk. There is no benefit to using the CACHE qualifier, and if used, would actually waste cache memory resources. |
$ MOUNT/CLUSTER/NOCACHE MDB150 FASTRAMDISK |
The DECram disk then can be accessed by any instance that is attached to the shared disk region.
$ MOUNT/NOCACHE MDB150 VOLUME-LABEL |
The command in this example mounts the shared device MDB150. If you want to share the RAM disk, then you must specify the /SYSTEM or /CLUSTER qualifiers.
Data on the shared DECram disk is preserved if you dismount and then mount the disk. In addition, data is also preserved on the DECram disk even if the Galaxy instance crashes, as long as another instance is attached to the shared region. |
See the OpenVMS System Management Utilities Reference Manual for more information on the MOUNT command.
2.3 Restoring an OpenVMS Alpha Version 7.2-1H1 (or Later) DECram Disk
Following are the three ways to restore all RAM disks:
The MDRECOVER.EXE image restores memory-based disks after a power failure or system crash. MDRECOVER.EXE uses memory disk data in SYS$COMMON:[SYSMGR]MDRECOVER.DAT to recover memory disks with the /PERSIST qualifier set.
Separate MDRECOVER.DAT files may not have the exact same list of disks in them and that a recovery from one node may not restore the disks of another cluster node. A common MDRECOVER.DAT file is required for consistent cluster-wide memory disk recovery. This file is invoked automatically when the system is rebooted. |
Issuing the "PRODUCT INSTALL DECRAM" statement at the DCL prompt invokes the POLYCENTER Software Installation procedure. POLYCENTER then installs the following command in the MDRECOVER.DAT file for execution by OpenVMS at system startup time:
$ MCR SYSMAN STARTUP ADD FILE MDRECOVER.EXE/PHASE=LPMAIN/MODE=DIRECT |
This command is issued automatically during installation and results in
the restoration of the RAM disk(s) automatically after a power failure,
system maintenance, or interruption of service for any reason.
2.3.2 Restoring a Disk Using the DCL command DECRAM RECOVER
The following command can be issued to recover a DECram disk:
$ DECRAM RECOVER |
The DECram command RECOVER reads from the MDRECOVER.DAT file located in the SYS$COMMON:[SYSMGR] area and restores each disk in the file that has the /PERSIST qualifier set.
The following DECram command recovers all the disks with /PERSIST set:
$ DECRAM RECOVER |
2.4 Determining Allocation of DECram Device Resources
The SHOW DEVICE command is used to show the resources that are allocated for existing RAM disks.
The DECRAM SHOW DISK command shows all of the disk resources that will be allocated on the next disk recovery when using the SYS$COMMON:[SYSMGR]MDRECOVER.DAT file. That command includes all disks with or without the qualifier /PERSIST being set.
If MDRECOVER.DAT is a cluster wide common file, it will not list local disks of other cluster nodes with different allocation classes. |
2.4.1 Determining Resource Allocation Using DECram Version 3.0 and Later
When using DECram Version 3.0 and later, you can display information
about a virtual disk from both the DECRAM$RECOVER.DAT and MDRECOVER.DAT
recovery files by issuing the DCL command DECRAM SHOW DISK.
$ DECRAM SHOW DISK [device-name] |
device-name
The device name used in the CREATE DISK command (see Section 2.2.2.2). If no device name is specified, a brief status is displayed for all disks in the recovery files.
When the command includes a disk specification, the following information is displayed about the disk:
When no device name is specified, only the disk name, size, and volume label (if available) is displayed for all disks in the recovery files.
If no output is displayed, no DECRAM$RECOVER.DAT or MDRECOVER.DAT files exist.
Reference the online help information available through the OpenVMS DCL HELP command or see the OpenVMS DCL Dictionary for more information on the SHOW DISK command.
$ DECRAM SHOW DISK |
This command does not specify a device, so a brief summary of information is displayed for all devices in the DECRAM$RECOVER.DAT and MDRECOVER.DAT recovery files.
From SYS$COMMON:[SYSMGR]MDRECOVER.DAT Disk $4$MDA0 Size 110 Label MDA0 Disk $4$MDF0 Size 100 Label MDF0MDF0 Disk $4$MDF1 Size 100 Label MDF1MDF1 Disk $4$MDF2 Size 100 Label MDF2MDF2 Disk $99$MDE0 Size 100 Label MDE0 |
$ SHOW DISK MDA0 From SYS$COMMON:[SYSMGR]MDRECOVER.DAT Disk $4$MDA0 Size 110 Label MDA0 Local memory, persist, serve, writebuffered. |
This command displays full information about device MDA0.
Never delete the MDRECOVER.DAT file. If you delete the file, then you will not be able to recover any disks after a power failure. |
The algorithm used by shadowing to select shadow set members for an IO operation is based primarily on member queue lengths. Although the queue lengths of all shadow set members will almost always be the same, DECram is expected to do more IO since it returns reads and writes much more quickly than a physical disk. One way to ensure that the application reads only from shadow set DECram disks is to do IO local to the DECram disk while the other shadow set members are served.
If you use the INIT/SHADOW DCL command to initialize a DECram disk followed by a MOUNT/SHADOW command, be aware that OpenVMS volume shadowing will not execute a full copy operation. To execute a full copy operation of shadowed disks and make the data fully consistent between shadow set members, you should issue an INIT/ERASE command after the INIT/SHADOW command. |
2.5.1 Mounting a Shadowed DECram Disk--DECram Disk to DECram Disk
To mount a shadowed DECram disk, use the following format with the DCL
command MOUNT for single or clustered systems:
$ MOUNT DScuuuuu/shadow=($n$device-name:) volume-label |
$ MOUNT/SYSTEM/NOCACHE DSA0: /SHADOW=($3$MDB100, $3$MDA0) FASTRAMDISK |
$ MOUNT/CLUSTER/NOCACHE DSA0: /SHADOW=($3$MDB100, $2$MDA0) FASTRAMDISK |
where:
DS
Represents the controller code identifier, which is always DS for a shadowed disk.c
Represents any single letter, A to Z.uuuuu
Represents any numeric value between 0 and 32,767.n
Represents the allocation class for this OpenVMS system.device-name
The full device-name returned by SHOW DEVICE device-name (see Section 2.2.1).volume-label
The unique disk label used in the INITIALIZE command.
$ MOUNT DSA0/NOCACHE/SHADOW=($2$MDA1000:) FASTRAMDISK |
Never issue a MOUNT command without the /NOCACHE qualifier. |
The command in this example mounts the host-based shadow set DSA0 with the single member $2$MDA1000.
The volume label is FASTRAMDISK.
Data on the shadowed DECram disk is preserved if you dismount and then mount the host-based shadow set. |
See the manual Volume Shadowing for OpenVMS for more information on allocation class
and the MOUNT command. OpenVMS System Management Utilities Reference Manual: M--Z contains additional information on
the MOUNT command.
2.5.2 Mounting a Shadowed DECram Disk--DECram Disk to Physical Disk or Partition
DECram can be used to shadow real physical devices or partitions provided there is enough physical memory available. You would do this in any situation where, in addition to data integrity, application speed and performance are also important.
For reads, the shadowing code will read from the DECram disk. Writes will be written to all devices and, therefore, will be much slower than a write to a DECram disk only. On power failure, the data will be safe because it is stored on physical media.
Volume Shadowing for OpenVMS does not currently freeze writes to the shadow set when the physical disk is no longer available. Therefore, be aware that in the current Version 3.0 release, if the physical disk fails, you will be writing information to a volatile disk and data may be permanently lost if the write is interrupted due to a power failure. |
If you have an existing shadow set that you know has more reads than
writes, placing a DECram disk in the shadow set can dramatically
improve overall performance. The reason for this performance
improvement is that a read operation does not require accessing a
physical disk because the information is available on the RAM disk.
Therefore, the time required to access the data is minimal. A write
operation in a shadow set comprised of a DECram disk still requires the
time needed to write the data out to the physical disk.
2.5.2.1 Creating the DECram Disk
Use the following command to determine the total number of disk blocks your physical device (or existing DSA device) has; this is the size you must use to create your DECram disk:
$ SHOW DEVICE/FULL $n$device-name: |
Use the DECRAM CREATE command to create your DECram disk with the /CAPACITY qualifier equal to total blocks. Then use the MOUNT command to add the DECram disk to the shadow set and initiate the shadow copy to the DECram disk. These commands are shown in the following example.
$ show dev dsa640/full Disk DSA640:,device type MSCP served SCSI disk, is online, mounted, file- oriented device, shareable, available to cluster, error logging is enabled. Error count 0 Operations completed 8107 Owner process "" Owner UIC [SYSTEM] Owner process ID 00000000 Dev Prot S:RWPL,O:RWPL,G:R,W Reference count 1 Default buffer size 512 ** Total blocks 1027362 Sectors per track 85 Total cylinders 756 Tracks per cylinder 16 Volume label "TST640" Relative volume number 0 Cluster size 3 Transaction count 1 Free blocks 1027203 Maximum files allowed 128420 Extend quantity 5 Mount count 2 Mount status System Cache name "_$84$DKC200:XQPCACHE" Extent cache size 64 Maximum blocks in extent cache 102720 File ID cache size 64 Blocks currently in extent cache 0 Quota cache size 0 Maximum buffers in FCP cache 4610 Volume owner UIC [SYSTEM] Vol Prot S:RWCD,O:RWCD,G:RWCD,W:RWCD Volume Status: subject to mount verification, file high-water marking, write-back caching enabled. Volume is also mounted on CSG6. Disk $4$DUA640:, device type MSCP served SCSI disk, is online, member of shadow set DSA640:, error logging is enabled. Error count 0 Shadow member operation count 8126 Host name "HSJ50R" Host type, avail HSJ5, yes Alternate host name "HSJ50G" Alt. type, avail HSJ5, yes Allocation class 4 Disk $4$DUA642:, device type MSCP served SCSI disk, is online, member of shadow set DSA640:, error logging is enabled. Error count 0 Shadow member operation count 8134 Host name "HSJ50R" Host type, avail HSJ5, yes Alternate host name "HSJ50G" Alt. type, avail HSJ5, yes Allocation class 4 $ DECRAM CREATE DISK MDA1642/CAP=1027362/MEMORY=LOCAL/SERVE/ALLOCLASS=84 $ mount/cluster/nocache dsa640/shad=$84$MDA1642: TST640 %MOUNT-I-MOUNTED, TST640 mounted on _DSA640: %MOUNT-I-SHDWMEMCOPY, _$84$MDA1642: (CSG84) added to the shadow set with a copy operation %MOUNT-I-ISAMBR, _$4$DUA642: (HSJ50R) is a member of the shadow set %MOUNT-I-ISAMBR, _$4$DUA640: (HSJ50R) is a member of the shadow set $ show dev dsa640 Device Device Error Volume Free Trans Mnt Name Status Count Label Blocks Count Cnt DSA640: Mounted 0 TST640 1027203 1 5 $4$DUA640: (HSJ50R) ShadowSetMember 0 (member of DSA640:) $4$DUA642: (HSJ50R) ShadowSetMember 0 (member of DSA640:) $84$MDA1642: (CSG84) ShadowCopying 0 (copy trgt DSA640: 1% copied) $ |
To mount a shadowed DECram disk, use the following format with the MOUNT command, using the existing shadow set device name:
$ MOUNT DScuuuuu/shadow=($n$device-name:) volume-label |
where:
DS
Represents the controller code identifier, which is always DS for a shadowed disk.c
Represents any single letter, A to Z.uuuuu
Represents any numeric value between 0 and 32,767.n
Represents the allocation class for this OpenVMS Cluster member.device-name
The full device-name returned by SHOW DEVICE device-name (see Section 2.2.1).volume-label
The unique disk label used in the INITIALIZE command.
$ MOUNT DSA0/NOCACHE/SHADOW=($2$MDA1000:,$2$DUA546) FASTRAMDISK |
The command in this example mounts the host-based shadow set DSA0 with the members $2$MDA1000 and $2$DUA546. The volume label is FASTRAMDISK. See Section 2.5.2.2 for more information on mounting a DECram disk.
See the manual Volume Shadowing for OpenVMS for more information on allocation class and the MOUNT command. OpenVMS System Management Utilities Reference Manual: M--Z contains additional information on the MOUNT command.
Previous | Next | Contents | Index |