Questions regarding system operation are explained.
Q: Do I need to exit all applications to perform synchronization copying on the volume? If it is not necessary, what effects should I be aware of?
A
You do not need to exit the applications, but there could be some affect on the performance. If the application response is seriously affected, you can either suspend the synchronization copying, or adjust the load caused by the copying.
Interrupting the copying:
# sdxcopy -I -c class_name -v volume_name |
Resuming the copying:
# sdxcopy -B -c class_name -v volume_name |
Adjusting the copying load:
# sdxcopy -P -c class_name -v volume_name -e delay=delay_time |
Checking the copying process:
# sdxinfo -S -c class_name -o volume_name -e long |
See
Q: How do I change the volume size?
A
Volumes can be expanded with the sdxvolume -S command during the service. For details, see "1.3.7 Online Volume Expansion."
The procedure for changing the volume size while suspending access to the volume is as follows.
Exit the application that is accessing the volume
Backup the volume data as necessary.
Remove the volume.
Check if you can create a volume with the necessary size in the group in which the volume existed. You can check this as follows.
# sdxinfo -V -c class_name -o group_name |
The volume which indicates FREE in the STATUS field will be available. The BLOCKS field indicates the volume size. When there is more than one available volume, the maximum size you can create would be the size of the largest available volume.
When you cannot secure sufficient size within the group by performing procedure 4., backup a different volume and remove it, or create another group.
Specify a new size and create the volume. When creating the volume, specifying the same name as the volume you removed in procedure 3. will reduce the amount of information you have to input when changing the device definition information in the following procedure.
Update the device definition information such as /etc/vfstab, as necessary.
Restore the backup data you collected in procedure 2.
Q: I mistakenly updated the /etc/system file, and lost the definition for system disk mirroring. Now, I cannot boot the system. How do I restore the file?
A
The file contents at the point when the system disk mirroring was configured are saved in the following file.
/etc/opt/FJSVsdx/bkup/etc/systemTry booting the system by following the procedures described below.
From the OpenBoot ok prompt, use the boot -a command and boot the system as instructed.
For the name of system file, type as below.
/etc/opt/FJSVsdx/bkup/etc/systemFor the physical name of root device, type as below.
/pseudo/sfdsk@0:2,blkFor details on the boot(1M) command, see the Solaris manual.
Q: I cannot boot the system. How do I restore it?
A
Q: When an I/O error status is created in the test, how do I shorten the restoration time?
A
The sdxfix -D command is used for reading check for the restoration target disk, the execution time may be long corresponding to the disk capacity.
If you specify the -x NoRdchk option, the reading check will be skipped and shorten the execution time. The -x NoRdchk option is only available when the reading check is obviously not needed such as an I/O error status is created. For details, see "D.9 sdxfix - Restore a failed object."
Q: When parting a proxy volume in a shared class, I want to lock the proxy volume not to start up on certain nodes.
A
Before parting a proxy volume, on nodes where you want to restrain the proxy volume from startup, turn on the proxy volume lock mode with the sdxattr -V command. For details on the lock mode, see "Lock Volume" and "D.7 sdxattr - Set objects attributes."
Q: Is there a possibility that joined proxy volumes are parted automatically?
A
In the situation where the master volume has only one ACTIVE slice, if an I/O error occurs in that slice, the proxy volume in STOP or INVALID status that is joined with the master volume through soft copy is parted automatically.