02-04-2015 06:37 AM
Hi,
i'm posting here because i've a really strange issue with OPsCenter Version 7.5.0.6:
when i execute a standard report ("Client Summary Dashboard" for example), it shows that 4 clients (out of more than 100) never had a kbyte backup, but have a 100% rate success (in a monthy interval, that means more than 25 successful jobs).
Obviously, from BAR or catalog interface, i can see that i have 25 valid images (each image is more than 2GB).
As i've said, i've this wrong report on 2 clients (part of a group of 4 clustered servers, that's to say all of these clients are equal, same SO, same NB agent version), while on the other 2 ones everything is ok.
can you please give me some hints to solve this issue?
thank you very much,
regards
Alberto
02-04-2015 08:37 AM
It would help if you post some evidence, not quite sure what you're trying to explain,
02-05-2015 07:23 AM
Hi Riaan,
thank you for your reply.
you can see below a part of a report from OPSCenter ("Client Summary Dashboard"). it's about 2 client (standard backup policy, backup selection was about / on 2 linux server). backup was OK for both these clients, BUT OPSCenter report shows that NO Data (VOLUME KB column) for CLIENT_B. (in BAR console interface i can see there are actual data ready to be restored)
Please note that total number of successful job is 30 for both clients.
Client Name | NetBackup Version | Hardware | Operating System | Volume(KB) | Average Throughput (KB/sec) | Total Jobs | Number of Successful Jobs | Number of Partially Successful Jobs | Number of Failed Jobs | Number of Restore Jobs | Success Rate (%) | |
CLIENT_A | 7.5.0.6 | Linux | Linux |
|
2692.03 | 30 | 30 | 0 | 0 | 0 | 100 | |
CLIENT_B | 7.5.0.6 | Linux | Linux | 0 | 0 | 30 | 30 | 0 | 0 | 0 | 100 |
Regards
02-05-2015 07:57 AM
Hi,
What do you get if you run say "Jobs by Application" for only the affected clients? Does it also show ZERO bytes?
02-06-2015 02:04 AM
Hi Riaan,
yes, same behaviour here (moreover, this report shows that a lot of files were backed up):
Client Name | Policy Type | Volume(MB) | Number of Files |
CLIENT_A | Standard | 16.31 | 826,078 |
CLIENT_B | Standard | 0 | 589,528 |
Regards
02-06-2015 02:23 AM
Hi,
Ok we could dig into this a lot more but in the end I think it would require something to be fixed in your database. So rather open a support case, something seems to be wrong there.
02-06-2015 07:26 AM
Thank you for you support Riaan,
i'll open a support case,
regards.
02-16-2015 02:03 PM
Try running a custom report, historical, selecting just that client, and set Y1 axis to report on "Job Protected Size", "Total". Does that work?
If that looks right, then go back and change the Y1 axis to just "Job Size". See what the report comes up with.
02-17-2015 05:40 AM
Hi Ron,
thank you for your reply, here are what i got from the 2 reports:
1) "Job Protected Size" report: it shows jobs from 5GB (incremental bck) to 40GB (full bck)
2) "Job Size" report: it gives me a "no data available" warning.
regards,
Alberto
02-17-2015 07:11 AM
Based on your other posts, it looks like you are using Media Server Deduplication Pools, which has created problems with OpsCenter's original, default reports because, for some reason, OpsCenter's default reports are using the value in "Job Size", which is the number of kilobytes actually sent when factoring in deduplication, instead of the value in "Job Protected Size", which is the number of kilobytes in all the files before deduplication. I don't know why it still shows zero kilobytes when there were at least a few KB sent, but that's what it does.
The only way to fix this, from our standpoint, was to create our own custom reports similar to the default ones, and tell it to use the Job Protected Size. Then it matched the reports generated from the NetBackup Java GUI, and matched what we expected to see when comparing historical reports before we started using MSDP's.
02-17-2015 07:44 PM
Hi Ron,
That issue you're referring to relates to the domain_jobarchive.byteswritten and domain_jobarchive.presissize figures.
The Jobs size report incorrectly uses domain_jobarchive.byteswritten (deduped figure in the case of MSDP) which as you say messes up the figures when using MSDP (DD and Tape work fine). But in that case it would show an incorrect (lower) value, this case shows nothing.
That is concerning.
02-18-2015 06:52 AM
This problem is present in 7.5.0.7, too, as that is what we are using. I see a FEW clients that are reporting zero KB, but can see true numbers if I run reports using Job Protected Size.