Clone backup savesets directly to archive savesets?

[goony]

I have full backups of various systems on my adv_file device (i.e. disk storage).

Is there any way that I could select full backup savesets for particular clients and "clone" them into the Archive pool? (Note: I haven't actually tried this but I guess I should!) )

I'd like to skip the step of actually taking a system down for a period of time to do a full archive of the system since I already have a full backup of the client.

Thanks!

 

[NWNinja]

You should use mminfo for this and create a slash (/) separated list of ssid/cloneid to use with the nsrclone command. See the examples below, you should modify them to meet your needs but that should basicly be it. For example, if the clients full backups always go to a specific pool, you should of course use that in your query to make the search less extensive.

1)
mminfo -av -xc/ -r ssid,cloneid -q client=calvin,level=full,type=adv_file,copies=2 > ssid.tmp

2)
nsrclone -v -b pool -S -f ssid.tmp

Cheers!
Maverick

 

[605]

I do not know. However, it could be the usage of Archive pools require the Archive license. This means extra money.


I guess you only want to do this because you want to assign a longer retention policy to the clone media.

Well, since NW 7.3 you may assign a retention policy to any pool. So simply create a new pool and assign a long retention policy to it. Then clone the save sets as usual.

 

[goony]

NWNinja, it was the right idea but trying to clone "backups" to "archives" just doesn't seem to be in the cards:

[tt]# cat ssid.tmp
2012033304/1173172503

# nsrclone -v -b 'Archive' -S -f ssid.tmp
Automatically copying save sets(s) to other volume(s)

Starting cloning operation...
nsrclone: SYSTEM error: pool `Archive' is not a cloning pool
nsrclone: Cannot open nsrclone session with mailserv1
nsrclone: Cannot open nsrclone session with mailserv1. Error is 'pool `Archive' is not a cloning pool'
nsrclone: Failed to clone any save sets

# nsrclone -v -b 'Archive Clone' -S -f ssid.tmp
Automatically copying save sets(s) to other volume(s)

Starting cloning operation...
nsrclone: SYSTEM error: Can only clone backup data to backup clone pools.
nsrclone: Cannot open nsrclone session with mailserv1
nsrclone: Cannot open nsrclone session with mailserv1. Error is 'Can only clone backup data to backup clone pools.'
nsrclone: Failed to clone any save sets
# [/tt]

I'm still fiddling around... it was a nice idea to not have to have a separate archive run for each client due to the following reasoning:

From what I can gather from the docs, it seems scheduled "Archive Requests" have no awareness of a performing a client shutdown/startup script ala savepnpc. In order to make a proper archive of a system with a live application (such as a database) I have to manually ensure that the database is in a quiescent state before beginning an archive.

 

[goony]

EMC/Legato support told me "you can't do what you want" so I'm now trying to think outside the box - any comments on either of these approaches (on my Unix server)?:

1) Some kind of scanner-output-piped-to-nsrarchive-input

For each desired saveset:
[tt]scanner -S <savesetid/cloneid> | uasm -rv | <some nsrarchive magic>[/tt]

I'm trying to figure out if there is a syntax for nsrarchive that would be valid for this approach.

2)Use a large space on the server to upack savesets, then nsrarchive to store the files.

The directory name holding the unpacked savesets would be something like /archive/<client>_<year>_<month>_<day>. When doing the recovery from the archive (initiated on the client with nwrecover) you would specify the name of the server instead of the real client, specify the archive indexes, find/select the files you want, and specify where to relocate them.

 

[605]

I don't understand where the problem is. In general, you do not want to do anything else but cloning a certain set of save sets to another pool's media. Sure you can do this.

In this case, ANY clone pool will do the job - it must not necessarily be an "archive" pool. This pool is for a first backup (see below). Why do you insist on this specific type?

The archive pool is still there for compatibility reasons but the original idea about archiving (backup without writing CFI info + plus true verification + deleting the original files (a move instead of a copy)) does not really make sense any longer as disks today have much higher capacity and only cost a fraction.

And ater the data has been written, NW does not care about the pool any longer. It will only look for the media the data has been written to.


So - let NetWorker work for you and do not force him to follow a scheme you have in mind just because it is there.

 

[goony]

I'll admit, I've had a fixation about something being an actual "archive" compared to just using a different clone pool and having an insanely long retention period.

In part, my fixation has been based on having an archive operation be able to "groom" (remove unwanted files) and do a verification of data written all in single operation is certainly attractive, but perhaps unnecessary given a little thought and some well-written scripts.

Thank you for taking the time to respond to my ramblings.

 

[NWNinja]

Sorry, i missed the part with the pool being an Archive pool.
So no, you cannot clone to a pool that is not a pool of type Backup Clone. However, i agree with 605, why do you need to keep the backups in the Archive pool? I would suggest you perhaps create a new pool which you call MyArchive or whatever. This should be a Backup Clone pool. Next, do the steps you did before with the cloning, only add an extra step with which you change the savesets browse/retention to "forever". Something like below.

for /f %i in (ssid.tmp) do nsrmm -w forever -e forever -S %i

Please note that if you are not using 7.3, the status of All copies will be changed! Cloneid is Not respected in pre 7.3.

Cheers!
Maverick