Attachment uploads are currently disabled. This is a temporary situation and will resume as normal in the coming days.
UNSOLVED
JO
josevtome-oxPni
1 Rookie
•
10 Posts
0
9328
February 10th, 2009 14:00
Networker server is not mounting tapes automatically with most backups
I have this issue with Networker 7.3.3.
I have a big infrastructure, but there are two servers that are having the same problem. The configuration is the following:
Operating System: Solaris 9
Networker Version: 7.3.3
Backup Library: StorageTek SL8500
ACSLS Software available for both sites.
Tape devices DA with networker server and a couple of Storage Nodes in one site.
Backup Types:
- Filesystem level backups (for application, home directories, and OS filesystems)
- Database: Oracle backups using RMAN.
The Oracle backups are not launched within Networker, but from crontabs or using ControlM application, so in most cases we have configured just a pool for the DB backup, with no client nor group (not required as RMAN only uses the pool).
The problem is that in some cases Networker is unable to mount the tapes automatically, and we have to mount the tapes manually either with NMC or command line. With the manual operation it works fine, but can't understand why It's not working automatically and it's driving us crazy.
This situation is most common for Oracle backups (either online or archive backups), but we can't determine if it happens always with X pool because sometime it works and sometimes it doesn't.
Anybody have an idea? Please we really need help because the system is killing us at nights! Thanks!
I have a big infrastructure, but there are two servers that are having the same problem. The configuration is the following:
Operating System: Solaris 9
Networker Version: 7.3.3
Backup Library: StorageTek SL8500
ACSLS Software available for both sites.
Tape devices DA with networker server and a couple of Storage Nodes in one site.
Backup Types:
- Filesystem level backups (for application, home directories, and OS filesystems)
- Database: Oracle backups using RMAN.
The Oracle backups are not launched within Networker, but from crontabs or using ControlM application, so in most cases we have configured just a pool for the DB backup, with no client nor group (not required as RMAN only uses the pool).
The problem is that in some cases Networker is unable to mount the tapes automatically, and we have to mount the tapes manually either with NMC or command line. With the manual operation it works fine, but can't understand why It's not working automatically and it's driving us crazy.
This situation is most common for Oracle backups (either online or archive backups), but we can't determine if it happens always with X pool because sometime it works and sometimes it doesn't.
Anybody have an idea? Please we really need help because the system is killing us at nights! Thanks!
- I guess you have pending message, is that correct? At that point, do you have appendable tapes in silo? If yes, do you see them via mminfo (check location field) and nsrjb? In daemon.log do you see request to load any tape? If yes, then can you check in acsls log did you had mount request for that tape?
josevtome-oxPni
1 Rookie
•
10 Posts
1792
0
Posted February 11th, 2009 06:00
Hi. Thank you for your reply. No, automedia management is not configured at the library. As I understand, correct me if I'm wrong, automedia management is used so the library can automaticaly label any blank tape for a pool needing one. This is not of my interest because the library labels all tapes without trying to relabel an expired tape first.
Do you consider that automedia management is required?- It looks like nsrjb command was done. This command should be passed by ssi on NW side to csi side on ACSLS side. So I would check ssi log and then you log on ACSLS side. Not sure of exact name now, but if you log a accss (or is it acssa?) then inside log directory (try to sort it by time stamp) you will find log which notes about CSI module tape requests and movements.
josevtome-oxPni
1 Rookie
•
10 Posts
1792
0
Posted February 11th, 2009 07:00
Hi. Yes, I have pending messager for operations waiting for tapes, the silo and specifically the pool waiting for tapes have available non full or expired tapes, some in undef state, others lower than 100% of data written (I can see them with a mminfo and nsrjb), the request for tape load is registered in the daemon.log The only thing I'm not sure is about the acsls log. In which log do you actually see a mount request?- I have or had a very similar issue but with NetWorker 7.4.2. When NetWorker got very busy it would not mount a tape from the request. The request would just sit there. I could do a manual mount of the tape which would work. I also discovered that I could just pick any tape in the library and do a inventory of it and check the box to "Reinitialize library state before inventory" and after the inventory, Networker would start mounting taes for it¿s request again, the issue would came back again later in a few hours. To help the issue, I spread out some of my backups and clone jobs so the system would not get so busy and that seems to help a lot. (I also had other strange intermittent errors) It¿s kind of stumped the EMC support folks. Since then I have upgrade to 7.5 and have not seen that problem again, yet anyway.
- Auto-media management is not a must. In any case, the order in which it picks tapes is
1. appendable tapes in the same pool
2. recyclable tapes that can be labeled to the required pool
3. new tapes
Enabling this is a matter of your convinience. I don't think it will solve your current problem though. Checking the logs for ACLS might give you a clearer picture of the problem. Increasing time outs might help too. josevtome-oxPni
1 Rookie
•
10 Posts
1237
0
Posted February 12th, 2009 07:00
The deposit timeout for my SL8500 s set to 15, should I change the value? any recommendation?
Hari5
443 Posts
1792
0
Posted February 10th, 2009 22:00