I have a filesystem that is too large to snap in 1 hour - which is the limit according to support. I need to get the filesystem backed up, because we are moving it to Isilon. There are 4 top level subfolders, and I am able to get 3 of them to finish. The last lab is just too big, and they have about 30 subfolders under their top level folder. I would like to split the folder up by using wildcards such as A-M and N-Z. I used to be able to do this with mailboxes in Exchange backups, and I'm hoping I can do the same with NDMP instead of listing out each subfolder in the save set field.
I assume since you mention Snapsure that this is a Celerra. You can use wildcards but must specify the complete path down to the directory you want to exclude. Following are the rules:
The path must start with a forward slash (/) and the file system name
The path can contain a maximum of 1,023 ASCII* characters
Supported wildcards are: question mark (?), which matches any single character and asterisk (*) which matches any sequence of zero or more characters
valid path= the complete path of the directory (and the specific type of subdirectories, if using wildcards) to exclude from the backup
For directory filtering, the question mark and asterisk wildcards can appear only at the end of the string
NOTE: The question mark wildcard is not accepted when Celerra international character sets are enabled. A message is sent to the server log when this is attempted.
The pattern-match algorithm is case-sensitive, by default
Example:
To exclude the directory /ufs1/dir1 from the NDMP backup, type:
EMC_EDIR01=/ufs1/dir1
To also exclude all directories that begin with "d" under /ufs2, type:
EMC_EDIR02=/ufs/d*
Based on your entry above it sounds like you want to exclude all directories under your mountpoint between let's say A-N. Let's say your mountpoint is /Users....so your entries would be something like the following:
EMC_EDIR01=/Users/A*
EMC_EDIR02=/Users/B*
EMC_EDIR03=/Users/C*
And so on up to EMC_EDIR50 (50 is the maximum based on default. It can be increased by modifying the filer.numDirFilter parameter)
Keep in mind it is case sensitive so if you might have some directories that start with A while others start with a....then you will need a separate entry for each one.
Last but not least.....please note that in spite of the filtering...you will still see the savesets appear to start but note that NOTHING will be backed up....I think the size will be something like 512KB or something which is basically just the NetWorker start mark for an NDMP backup.
If you are interested this is also documented in the "Configuring NDMP Backups on Celerra" which is available on Powerlink.
Thanks for the response. I actually need to do the opposite. I need to split the backup into smaller chunks, so I was hoping that I could do something like this for the saveset.
/root_vdm_3/group12/A*
/root_vdm_3/group12/C*
so that they would run separately or I could run 2 backups of the same group but with different subdirectories. There are about 20 directories under the root and I would rather not have to list them all.
This should work fine. Just keep in mind that it has to be:
1) The full path starting from the mountpoint down to the directory (so if you have /root_vdm_3/Users/group12/A*....or /root_vdm_3/group12/A*....but must be exact right down to that A*).
2) It is case sensitive
3) Maximum of fifty per backup
4) Each entry (directory path) is on a separate line in the Application Information section of the Client resource
5) Each entry increments the counter by 1....EMC_EDIR01, EMC_EDIR02, EMC_EDIR_03 and so on.
rmines1
49 Posts
2407
0
Posted February 21st, 2012 10:00
Hi Kathy:
I assume since you mention Snapsure that this is a Celerra. You can use wildcards but must specify the complete path down to the directory you want to exclude. Following are the rules:
The path must start with a forward slash (/) and the file system name
The path can contain a maximum of 1,023 ASCII* characters
Supported wildcards are: question mark (?), which matches any single character and asterisk (*) which matches any sequence of zero or more characters
valid path= the complete path of the directory (and the specific type of subdirectories, if using wildcards) to exclude from the backup
For directory filtering, the question mark and asterisk wildcards can appear only at the end of the string
NOTE: The question mark wildcard is not accepted when Celerra international character sets are enabled. A message is sent to the server log when this is attempted.
The pattern-match algorithm is case-sensitive, by default
Example:
To exclude the directory /ufs1/dir1 from the NDMP backup, type:
EMC_EDIR01=/ufs1/dir1
To also exclude all directories that begin with "d" under /ufs2, type:
EMC_EDIR02=/ufs/d*
Based on your entry above it sounds like you want to exclude all directories under your mountpoint between let's say A-N. Let's say your mountpoint is /Users....so your entries would be something like the following:
EMC_EDIR01=/Users/A*
EMC_EDIR02=/Users/B*
EMC_EDIR03=/Users/C*
And so on up to EMC_EDIR50 (50 is the maximum based on default. It can be increased by modifying the filer.numDirFilter parameter)
Keep in mind it is case sensitive so if you might have some directories that start with A while others start with a....then you will need a separate entry for each one.
Last but not least.....please note that in spite of the filtering...you will still see the savesets appear to start but note that NOTHING will be backed up....I think the size will be something like 512KB or something which is basically just the NetWorker start mark for an NDMP backup.
If you are interested this is also documented in the "Configuring NDMP Backups on Celerra" which is available on Powerlink.
Hope this helps.
Regards,
Rita Mines