PRIMECLUSTER Global File Services Configuration and Administration Guide 4.2 (Solaris(TM) Operating Environment)

F.2.18 sfcproxyjoin(1M) Sets up the relationship for high speed backup and restoration

[SYNOPSIS]

sfcproxyjoin equiv[alent] -r rel_name [-s] master_raw_device_file=proxy_raw_device_file

sfcproxyjoin equiv[alent] -r rel_name [-s] -f param_file

sfcproxyjoin equiv[alent] -r rel_name [-s] -g raw_device_file master_class.master_group=proxy_class.proxy_group

sfcproxyjoin equiv[alent] -r rel_name [-s] -g master_raw_device_file -f param_file

sfcproxyjoin separate -r rel_name [-s] master_raw_device_file=proxy_raw_device_file

sfcproxyjoin separate -r rel_name [-s] -f param_file

[DESCRIPTION]

sfcproxyjoin sets up the relationship between backup source and destination for high-speed backup and restoration. This function provides a snapshot that allows users to create easy a replica (snapshot) of a GFS Shared File System when the file system is mounted with PRIMECLSUTER Global Disk Service Snapshot function.

Specify a relationship name as the name that corresponds to the relationship. After the relationship is set up, execute all operations such as backup, restoration and relationship cancellation with the relation name specified.

In the case of EC operation, specify equivalent (or equiv) in the first argument. In this case, the backup source and destination enter the join status and copy from the backup source to the backup destination is started. When the -s option is not specified, sfcproxyjoin returns without waiting for the real copy processing to be completed. The file system of the backup destination that entered the join status and the partitions that compose that file system are disabled until they are separated by using sfcproxyop part, or the relationship is canceled by using sfcproxybreak(1M).

In the case of OPC operation, specify separate in the first argument. In this case, although One Pont Copy is performed from the backup source to the backup destination, sfcproxyjoin returns with the backup source and destination remained in separate status. The copy operation is performed logically instantaneously but file system update operation at the backup source is temporarily locked when a copy command is issued. When the -s option is not specified, sfcproxyjoin returns without waiting for the real copy processing to be completed.

When a relationship is set up with sfcproxyjoin, the partitions of the backup destination are set up for the corresponding partition that compose the GFS Shared File System of the backup source. In order to use the hardware copy function of EMC's Symmetrix storage device, a pair of groups instead of a pair of volumes must be set as a master and a proxy. In this case, the pair of the group of GDS including the GFS Shared File System and the group that is the backup destination must be specified with -g option.

And a relationship for GDS is also set up between the corresponding objects. The backup source is set as master, and the backup destination is set as proxy. Consequently, conditions of a relationship for PRIMECLSUTER Global Disk Service Snapshot need to be satisfied between the corresponding objects. Main conditions are shown below. Refer to "PRIMECLUSTER Global Disk Services Configuration and Administration Guide" for details.

Precondition of volume composition of GDS are shown below.

• Master and proxy belong to different group of same disk class of GDS.
• The master and proxy volumes must be the same size. In the case of group relationship, the proxy group must be greater than or equal to the master group size.
• In the case of group relationship, the proxy group must not have any volume.
• Volume being related as master is not related as proxy.
• Volume being related as proxy is not related as master or proxy.
• Although two or more backup destinations can be related to one backup source, different operation pattern cannot be mixed. For example, the file system already related as backup source for OPC operation cannot be related as backup source for EC operation.

Precondition to use hard copy function of storage device are shown below. When not satisfying it, the copy is done by GDS.

• The master and proxy volumes must not be stripe or concatenation-type volumes.
• The proxy must not be mirrored.
• When the OPC operation is performed, the volumes (that correspond to one another) between the backup source and backup destination exist on the Fujitsu RAID device ETERNUS6000 or ETERNUS3000.
• When the OPC operation is performed, the master and proxy volumes are not mirrored.
• When the EC operation is performed, the volumes (that correspond to one another) between the backup source and backup destination exist on the Fujitsu RAID device ETERNUS6000 or ETERNUS3000.
• When the EC operation is performed, the groups (that correspond to one another) between the backup source and backup destination that fulfills the conditions of EMC TimeFinder or EMC SRDF.

The file system of the backup source and destination need to satisfy the following conditions.

• The file system of the backup source must be a single partition.
• Do not exist a file system in the group except the file system of the backup source when you operate EC of each group.
• When the file system is already created at the backup destination, the configuration of MDS nodes and nodes that are sharing each file system for the backup source is the same as that of the backup destination.

When relationship is set up, sfcproxyjoin fails in the following cases.

• The case that backup source or destination volume of GDS is copy status.
  In this case, execute this command after the copy is completed.
• The case that the status of backup source volume of GDS is data invalid (INVALID) status.
  In this case, recover from INVALID status refer to "Appendix Resolving Problems" in the "PRIMECLUSTER Global Disk Services Configuration and Administration Guide."
• The case that the partitions of backup destination are activated on some nodes other than the node on which sfcproxyjoin is executed.
  In this case, inactivate the volumes by using sdxvolume(1M) of GDS on those nodes.

When a relationship is set up, a file system needs to have been created at the backup source but not at the backup destination.

[OPTIONS]

The following options can be specified:

-f param_file

Specify a file in which the correspondence between the backup source and destination is described. Describe the correspondence between partitions in the format "master_raw_device=proxy_raw_device" as in the case of the operand. Describe the correspondence between groups in the format "master_class.master_group=proxy_class.proxy_group." For details, see the OPERAND items. When a line begins with "#", it is assumed to be a comment line, and content up to the end of that line is ignored.

-g raw_device_file

Specify the relationship of group of GDS. The path name of a character special device of the representative partition of GFS Shared File System that is the backup source is specified for raw_device_file.

-r rel_name

Specify a character string of 255 or less characters as a name that is assigned to this relationship. However, these characters must not begin with "-".

-s

Synchronization specification. This option waits for the real copy processing to be completed.

[OPERAND]

Either of the following operands can be specified:

master_raw_device=proxy_raw_device

Specify the partition to be a backup source in master_raw_device and the one to be a backup destination in proxy_raw_device. These two partitions must be the same size.

In addition, it is effective only when -g option is not specified.

master_class.master_group=proxy_class.proxy_group

Specify the class and group of GDS that include the partition to be a backup source in master_class.master_group and the one to be a backup destination in proxy_class.proxy_group when -g option specified. The proxy group size must be larger than master group size.

In addition, the volumes of the backup destination group are automatically created by the name that adds "-" to the volume name of the volume which exists in the backup source group.

[EXAMPLES]

(Example 1)

Use a GFS Shared File System whose representative partion is /dev/sfdsk/class0001/rdsk/volume0001 as a backup source. To relate the partition /dev/sfdsk/class0001/rdsk/volume1001 as a backup destination of this file system for the EC operation, Execute the following command.

# sfcproxyjoin equivalent -r backup1_for_volume0001 \
/dev/sfdsk/class0001/rdsk/volume0001=/dev/sfdsk/class0001/rdsk/volume0001-

(Example 2)

Use a GFS Shared File System whose representative partion is /dev/sfdsk/class0002/rdsk/volume0002 as a backup source. To relate the partition /dev/sfdsk/class0002/rdsk/volume0002- as a backup destination of this file system for the OPC operation, Execute the following command.

# sfcproxyjoin separate -r backup1_for_volume0002 \
/dev/sfdsk/class0002/rdsk/volume0002=/dev/sfdsk/class0002/rdsk/volume0002-

(Example 3)

Use a GFS Shared File System whose representative partion is /dev/sfdsk/class0003/rdsk/volume0003 on group0003 of GDS as a backup source. To relate the group0003- of GDS as a backup destination of this file system for the EC operation, Execute the following command.

# sfcproxyjoin equivalent -r backup1_for_group0003 \
-g /dev/sfdsk/class0003/rdsk/volume0003 \
class0003.group0003=class0003.group0003-

[EXIT STATUS]

The following exit values are returned:

0

Successful completion.

>0

An error occurred.

[SEE ALSO]

sfcproxybreak(1M), sfcproxyinfo(1M), sfcproxyop(1M), sfcproxyrestore(1M)

sdxattr(1M), sdxinfo(1M), sdxproxy(1M), sdxvolume(1M) in the "PRIMECLUSTER Global Disk Services Configuration and Administration Guide."

[NOTES]

• Since full copy is performed after the relationship is set up, an I/O load is applied on the backup source and destination.
• Optional product PRIMECLUSTER Global Disk Services Snapshot is required for this function.