Using alternate media servers when it shouldn't

[PaulStevens]

I have an SLP that is using an alternate media server for the duplication step when it shouldn't, and I cannot for the life of me figure out why.

We use SSO (Shared Storage Option) for the tape drives and all the drives are zoned to master01, media02, and media03.  Each server has it's own stu for the tape drives.  master01-hcart-robot-tld0, media02-hcart-robot-tld0, media03-hcart-robot-tld0

All servers are Netbackup appliances with their own deduplication pools.  We also have an Access appliance that we use for longer on-site retention.  The Access is one big deduplication pool.  Each server has it own stu to talk to it, but they all target the same disk pool on the access.

The flow is the backup goes to the NBU appliance media server, then is duplicated to tape for long term retention, then duplicated to the Access for longer onsite retention.

SLP specifies media03 disk pool for the backup.   stu-disk-media03 is set to "on demand only" and "Only use the following media servers".  media03 is checked, media02 is unchecked. 

SLP specifies media03 tape drives for the first duplication step.  media03-hcart-robot-tld0 correctly has media03 as the media server and is set to "on demand only"

SLP specifies media03's connection to the access appliance for the second duplication step.  media03-acc99-dedup is set to "on demand only" and "Only use the following media servers".  media03 is checked, media02 is unchecked. 

The issue is happening on the tape duplication step between the media server and the tape drives, so the Access should have nothing to do with it.

Below is a snippet from the job log showing that it is grabbing the other media server.

Any suggestions what else to look at?

Feb 22, 2023 7:00:16 AM - requesting resource  LCM_media03-hcart-robot-tld-0
Feb 22, 2023 7:00:16 AM - granted resource  LCM_media03-hcart-robot-tld-0
Feb 22, 2023 7:00:16 AM - started process RUNCMD (pid=222774)
Feb 22, 2023 7:00:17 AM - begin Duplicate
Feb 22, 2023 7:00:17 AM - requesting resource  media03-hcart-robot-tld-0
Feb 22, 2023 7:00:17 AM - requesting resource  @aaaag
Feb 22, 2023 7:00:17 AM - reserving resource @aaaag
Feb 22, 2023 7:00:17 AM - ended process 0 (pid=222774)
Feb 22, 2023 7:00:19 AM - awaiting resource media03-hcart-robot-tld-0. No drives are available.
Feb 22, 2023 7:38:34 AM - Waiting for scan drive stop TLD000-D13, Media server: media03
Feb 22, 2023 7:38:37 AM - Info bpduplicate (pid=222774) window close behavior: Suspend
Feb 22, 2023 7:38:37 AM - resource @aaaag reserved
Feb 22, 2023 7:38:37 AM - granted resource  Y07697
Feb 22, 2023 7:38:37 AM - granted resource  TLD000-D13
Feb 22, 2023 7:38:37 AM - granted resource  media03-hcart-robot-tld-0
Feb 22, 2023 7:38:37 AM - granted resource  MediaID=@aaaag;DiskVolume=PureDiskVolume;DiskPool=dp_disk_media03;Path=PureDiskVolume;StorageServer=media03;MediaServer=media02
Feb 22, 2023 7:42:08 AM - Info bptm (pid=1009) start

[PaulStevens]

all Netbackup servers on version 9.1.0.1

[davidmoline · Accepted solution]

Hi @PaulStevens 

I'm sure the issue is that the disk pool (or more precisely the storage server) on media03 is set to allow access to media02 as well. Assuming this is the case, the NetBackup can choose either media server to read the backup image for duplication (the STU used has no bearing on reading the backup image). If you wish to force it to use media03 (which makes sense since that is where everything is for the operation), use the alternate read server in the SLP operation and select media03. 

If you don't need to duplicate between MSDP on media02 or media03, you could remove the access for media02 to the media03 stosrage server (accessed via the Credentials -> Storage Servers in the GUI).

Cheers
David