NBU 10x tape 2 tape copy (inline copy) clarification.

[KBfinder506]

Hi people:

We have a 4 drive LTO9 library, Master/media server is Windows2022 and NBU 10.1.2. I have some policies that do a two tape copy of data to different pools. For example, mydata policy have two schedules

foreverfull schedule make a backup to Mypool01 (primary copy on tape1) and its copy is located in Mypool02 (second copy on tape2) with retention level 100. Most of the time the policy make the two tape backup and copy successfully.

monthlyfull schedule make a backup to Monthpool01 (primary copy) and its copy is located in Monthpool02 with retention level 10.

Sometimes tape1 o tape2 fails over the copy  (HW errors mostly, like need cleans drive) and then just use bpduplicate to make the copy over the tape of the respective pool.

Today I have some doubts about the correct syntax, because I need to copy a backup with two copies made with monthlyfull schedule (the backupid is server01.net_1705357985)  to use foreverfull pools and change the retention level.

My question/doubts are around the -npc; should I use it? Of course, I want that backups on foreverfull schedule  remains as primary copies.

bpduplicate -v -number_copies 2 -backupid server01.net_1705357985 -client server01.net -dp Mypool1,Mypool2 -dstunit mylibrary-01-hcart3-robot-tld-0 -id 000012 -L copi.wri -policy Respaldo_servr01 -rl 100,100

Can anybody give some ligth about this?

[StefanosM]

The command you provide will create two new copies.

I assume that you want to create one copy of the backup with the backupID server01.net_1705357985. The simplest command that does the job is (-v and -L are optional but recommended)

bpduplicate -v -backupid server01.net_1705357985 -dp Mypoolx -dstunit mylibrary-01-hcart3-robot-tld-0 -L copi.wri -rl 100

All other parameters are not needed. You may use them as search criteria if you do not use backupid.

If you need to create two copies from a backupid then use

bpduplicate -v -number_copies 2 -backupid server01.net_1705357985  -dp Mypool1,Mypool2 -dstunit mylibrary-01-hcart3-robot-tld-0,mylibrary-01-hcart3-robot-tld-0 -L copi.wri  -rl 100,100

-npc is to change the primary copy, and it is not essential. If I remember correctly, you cannot use it with other paramiters but only with backupid to change the primary copy, and it it does not do duplications. To change the primary copy during duplication, use -set_primary.
If you omit set_primary, then the infinity backup will become primary when the first copy expires.

Last, you can not have two primary copies. Only one copy is the primary and is used for restores.

[KBfinder506]

Thanks @StefanosM , I will try it and give the community feedback.

[KBfinder506]

Hi people at first try

Activity monitor job id = 1189
Copy 1 activity monitor job id = 1190
Copy 2 activity monitor job id = 1191
The number of specified storage units must be equal to the number of copies

So weirdly have to specify the same -dstunit mylibrary-01-hcart3-robot-tld-0 twice separated with comas.

[KBfinder506]

Hi people, 3rd try

Activity monitor job id = 1195
Copy 1 activity monitor job id = 1196
Copy 2 activity monitor job id = 1197
INF - Skipping copy 2 of backup id server01.net_1705357985, is not required copy 1.

All jobs are queued since NBU says there are no drives available. But on NBU just two driver (from a previous backup) are used. The same says the library remote panel two drives on use, two empty. Yes the dsunit is configured as robot number: 0 and have on 4 the Maximum concurrent write drives.

[StefanosM]

are you sure that all drive have the same density type?

When all backups, restores, and duplications are finished, run the following command:
/usr/openv/netbackup/bin/admincmd/nbrbutil -dump

and post the lines, starting with
MdsAllocation

If no backups, restores, or duplications are running, there must be no lines with MdsAllocation.
If there are lines , run the command.

/usr/openv/netbackup/bin/admincmd/nbrbutil -resetall

[KBfinder506]

Hi people!

Dear @StefanosM yes sir they are the same type/manufacture/firmware and a long etc.

I'm attaching some captures.

Sadly, I can't run that command until a very long backup finished.

I feel that I'm loosing some detail with this procedure.

Since the two tape copy failed, I try to take the copy with monthly schedule (is on tape14) and copy with different retention level and different pool server01.net_1705357985. The command finished fine, no warning. That monthly backup expired today at 1am. But the command was run yesterday (04/19) at 3pm and finished today at 1:45 am.

Do I have to make something in addition to make that new copy located on tape15 be the copy1 ? I run the GUI and with image on tape report on tape15 appear my copy but with retention 1 month. Seems that -rl parameter was not applied.

To make things crazy, I just have run

bpduplicate -v -number_copies 1 -backupid server01.net_1705357985 -client server01.net -dp Mypool2 -dstunit mylibrary-01-hcart3-robot-tld-0 -id 000015 -L copi.wri -policy Respaldo_servr01 -rl 100
Activity monitor job id = 1214
INF - found no images or media matching the selection criteria

I'm dismayed about this. In my mind;  I can use the new copy located on tape 000015 and make the 2nd copy that will live on Mypool2 and tape ADR015. But find that messages and the image on tape report; says that backupd id has a retention of 1 month.

I'm really sorry to bother the community but I feel a bit lost. My NBU experience was installing and some GUI management... I can't call myself a backup operator.  Can somebody shed some light about this?

For the record, the version on NBU is 10.3 build 0062.

As you may guest, tapes on Mypool1 are labeled 0000XY, but tapes on Mypool2 are labeled as ADR0XY.

My other successful copies were made to comply with the rule of two tapes. But they were from the same schedule and retention level; just from Mypool1 to Mypool2 or viceversa. This copy is special because came from different schedule, retention time and pool (Monthpool1).

[KBfinder506]

Almost forgot the screen captures.

[StefanosM]

Run the bpimagelist command to see if the backupid is in the netbackup database, how many copies are and  where they are .
bpimagelist -backupid server01.net_1705357985

Again, you do not have to add so many arguments in the bpduplicate command. Keep it simple.
The folowing command will use the primary copy to create another one.
bpduplicate -v -backupid server01.net_1705357985 -dp Mypool2 -dstunit mylibrary-01-hcart3-robot-tld-0 -L copi.wri -rl 100

You cannot change the copy number of a backup copy. Only to make it primary. use
bpduplicate -npc x -backupid server01.net_1705357985

I suggest you use SLPs to create the two-copy strategy. Using the two-copy configuration in the schedule will lead to missed copies if a copy fails. With SLPs, no matter what, you will have two copies automatically.