Netbackup 7 concurrent jobs issue

[Hostasaurus]

Hi all, I've got an issue I'm trying to figure out on NB 7.0.1.  I've got one media server that has both a tape library and a large storage array attached to it.  The storage array is configured as a PureDisk disk pool.  The maximum concurrent jobs setting is set at 15 currently.  The tape library is set to 2 concurrent write drives and 15 streams per drive with 'enable multiplexing' checked.  The master server -> global attributes -> maximum jobs per client is set to 2.

I have each client in its own backup policy because we like to stagger full backups across the week and may need to adjust clients to run, not run, differnet start times, etc.  No policy has the "limit jobs per policy" check box on.  No policy has the "allow multiple data streams" check box on.  Each client's policy has a daily backup schedule on six days/week, and a full backup scheudle occuring on that client's day of the week for full backup.  In the full backup schedule I have media multiplexing set to 15 and 'multiple copies' is checked with the primary copy 1 going to puredisk and copy 2 going to tape.

So the problem I'm running into is disk-based daily backups seem to work fine; they kick off and 15 jobs stay running simultaneously until all the policies have completed.  However, the full backup jobs kick off and only back up one client at a time.  I see each policy get a job queued up, that turns into four total jobs where one appears to be the parent, one is a dedupe job, one is a disk-based backup job and one is the tape-based backup job.  All the other client jobs sit around until each can run one at a time.  The status reported in activity monitor is:

"Drives are in use in storage unit (backup5-hcart2-robot-tld-0)"

or if I view details:

"awaiting resource backup5-hcart2-robot-tld-0. No drives are available."

So it sounds like Netbackup is ignoring my setting of 15 for the 15 streams per drive multiplexing?  I've used the "bpsturep" command to set the Max MPX on the puredisk volume to 15 in case that somehow plays a role when you're doing a multi-destination backup that involves disk and tape but that did not seem to make any difference.

[Andy_Welburn]

"four total jobs where one appears to be the parent, one is a dedupe job, one is a disk-based backup job and one is the tape-based backup job"

how it's set up & the sequence?

The only thing that was crossing my mind was if duplications were being carried out from your tape images thereby tying up your media/drives?

[Hostasaurus]

Each policy has a full backup schedule, and that particular schedule has the multiple copies option checked with disk being primary, tape being secondary.  In activity monitor, the first of the four jobs shows up with "-" for the schedule name, no storage unit first, and the same id number as itself in the parent id column, 1106 in the case of the one I ran last night for testing.  The second job shows up with job id 1107, the correct schedule name, no storage unit, and the parent id set to 1106.  Now, the third and fourth jobs, they both have a parent id set to 1107, correct schedule and then the one has the dedupe storage unit and one has the tape storage unit, same bytes backed up, etc.

[J_H_Is_gone]

your error shows a specifice storage unit which also has a setting.

Your schedule says 15 jobs for multiplexing.

But look at the config of the storage unit being used by the policies.

in there you can alos set the max drive and the mulitplexing.

My first thought is that the storage unit is set to one, were your policies is set to 15.

Storage unit wins and only does one.

[Hostasaurus]

I went back to verify this and the storage unit setting is also at 15; that's what I meant by "The tape library is set to 2 concurrent write drives and 15 streams per drive with 'enable multiplexing' checked." in my first post.  I think that's the right place; in the Java GUI I went to:

NetBackup Management -> Storage -> Storage Units -> backup5-hcart2-robot-tld-0 and its properties show enable multiplexing, and maximum streams per drive set to 15.  If I run bpstulist -U I get:

Label:                backup5-hcart2-robot-tld-0
Storage Unit Type:    Media Manager
Host Connection:      backup5
Number of Drives:     15
On Demand Only:       no
Max MPX/drive:        15
Density:              hcart2 - 1/2 Inch Cartridge 2
Robot Type/Number:    TLD / 0
Max Fragment Size:    1048575 MB

Just in case, I checked under Media & Device Management -> Devices -> Drives and the properties of the individual dirves but there's nothing related to concurrency there.  Same with Robots.

[Baif]

1. Log Service Request to Symantec NetBackup Tech Support
2. Ask them tell you how to enable nbemm & nbrb logs to level 6.
3. Run test, and check nbemm & nbrb logs.
4. Ask them verify logs, keyword "EMMProvider" in nbrb debug logs.
5. Ask them cut the craps about EMM database, re-configuring STU etc, and give EEB to fix the problem
6. Ask sale from Symantec to give you free upgrade to 7.1.

FYI:

vxlogcfg -a -p NB -o nbemm -s DiagnosticLevel=6 -s DebugLevel=6

vxlogcfg -a -p NB -o nbrb -s DiagnosticLevel=6 -s DebugLevel=6

vxlogview -o nbrb --tail 00:15:00

vxlogview -o nbemm--tail 00:15:00