How Netbackup Command works

[Vickie]

Dear experts,

I have a query & concern regarding the COMMAND we use in NETBACKUP.

I wonder & doubt as from where and how these COMMANDS get the output, 

Lets take an eg,

Command,

./bppllist ( will give list of policy ). 

How it get the list, first where it will go or hit to get policy name. 

What is the exact route or way it'll follow to give the output.

Whom it'll ask or query for the list.

Same query is for every COMMAND we use in NETBACKUP,

bpplclients,

bpplinfo,

bpclient,

bpimagelist,

vmquery.

vmupdate, and many more....

Please assist me, if this is possible we get the way as how command works.

[Dyneshia]

The commands guide gives you all this information :

NBU 7.1 - NetBackup Commands Reference Guide  http://www.symantec.com/docs/DOC3684

NBU 7.5 - NetBackup Commands Reference Guide   http://www.symantec.com/docs/DOC5182

It will show you what it does, what parameters a needed , examples, and where to run it from.

[Vickie]

That's fine Dyneshia.

But my cocern is " From where these commands fetch the information", must be from EMM db

.

Which daemon or process these command query to find the relative information.

For example "We ran ( bpplclients), it'll give list of client exist in policy", but what exactly the path/way/method it follow to get the client list.

[Dyneshia]

Please reivew the SYSADMIN guide and review : http://www.symantec.com/docs/DOC5157

Parts of the NetBackup catalog - p780

About the NetBackup image database- p782

About the NetBackup relational database - p784

About the Enterprise Media Manager (EMM) database - p785

[mph999 · Accepted solution]

In 7.5 most information comes from the NBDB database. Policy info still comes from flat files, and part of the catalog db comes from the .f files (binaries, but is converted). There is little chance of finding documentaion on the exact path each command uses : 1. It doesn't exist, well not for every command 2. Even Tech Support doesn't have this (we have to work it out ourselves ) 3. Generally, you don't need to know the exact path they take Generally, Commands in volmgr/bin get their info from EMM (apart from scsi_command and scan) nbemmcmd command from emm db bplist/ bpflist from .f files in catalog bpimmedia / bpmedialist /bpimage from EMM nbstlutil from EMM That covers most of the day to day stuff I think. If you want to find out exactly how : 1. Turn on all the logs (yep, every single one) 2. Run a command, check which logs changed 3. Uses these logs, find the lines and put them in time order 4. Delete the lines that aren't relevant You have the answer ... There is a big problem with this ... Some logs update constantly, even when the system is idle. (EMM / bpdbm / nbrb / nbstserv to name a few) Therefore when you run a command (eg. nbemmcmd -some options) some of the lines in EMM will be due to the command, but some will be due to it just doing it's normal day-to-day things. It can be worked out, by finding the PID/ TID, but then again, EMM is multi threaded and so even these can change. Some commands jump from log to log - for example, many command in netbackup/bin/admincmd start with entries in the admin log, but this might go out and make a connection via PBX to another process, which logs in another log, but then again, the connection might go via vnetd, or perhaps nbproxy gets involved ... you can bgin to see the problem. How do I do it, well a combination of the above generally - I have a script that will tell me what legacy logs change when I run a command which helps, but is in no way the complete answer. Depending on what I'm tracing (and it's usually some job, so multiple processes / commands involved) it can take hours, and if I make a mistake, or miss something, I have to start again. There are flowcharts (trouble shooting guide) that show how some things work (eg. backups / restores) and there are some TNs / Internal only documentaion for others but not everything is covered. It's pretty much impossible to work out every command, and even if you did, as soon as a new version of NBU is released you can be pretty sure something somewhere changed, so youd have to start again. I, and I suspect my colleagues work on a need to know basis. If we need that level of detail, we work it out. Generally with experience, you have at least an approimate idea, and very often this is sufficient. You would be much much better off spending your time doing the following. Learning how to collect the logs requested by tech support (many people can't do this). Learning how to narrow down an issue through testing Learning when to recognise that NBU probably isn't the cause, or the main cause (hint, network issues and drive/ library issues) Troubleshooting techniques > start with the most likely cause first (yes you might be wrong, but on 'average' this is a winning approach) Understanding that when you find a problem that might be the cause of the symptoms, this HAS to be eliminated before further troubleshooting takes place. Knowing the importance of setting the PATHS for NBU commands in the OS PATH variable Martin

[Nicolai]

Wow - Martin post was really impressive.

My humble to comment.

How a commercial software vendor write code is something they keep to them self. If you really want to see how code work look at open source code like Apache,Chrome or even Amanda (backup utility)

Netbackup also contain open source software, and Symantec distribute this code according to GNU

See: http://www.symantec.com/docs/TECH52575

Last call is debugging the code using gdb - but that is really hard unless you a code expert.

[mph999]

Thank you Nicolai, for your kind comment.

It is true that a company will not release the code, but I don't think 'Netbackup_user' was asking that.   I think he was after 'flowcharts' level of info, but unforntunately, this does not exist for all processes / commands.

M

[Vickie]

Thanks All,

Thanks to you Martin, to write such a nice comment which makes me understand as approx idea is sufficient instead of going in deep flow of work for command atleast.

At the last in your comment, is quite impressive & be grateful for.

[mph999]

No problem, pleased to help.

[Vickie]

Martin / All,

One more thing I would like to ask if you spot some light on it as well. Please brief me regarding logs, which logs to follow for which sort of error or issue.

Thanks

[revarooo]

Netbackup user. It really depends on the issue. There are many logs and many issues that can arise that will change which logs you would need to look at.

Have a look at these Technotes which give some detail on old legacy and new unified logs

http://www.symantec.com/business/support/index?page=content&id=TECH18012

and this

http://www.symantec.com/business/support/index?page=content&id=TECH43251

When you enable the logging, just enable it for whichever part you need. Enabling logs will use up considerable amounts of disk space, especially for unified processes (nbemm,nbjm, nbpem and nbrb) and also bpdbm on the Master. If you increase the log levels they will grow even bigger and may use up all your free disk space. A full disk (if your logs are hosted on the same filesystem/disk as your NetBackup installation) can cause NetBackup to shut itself down to preserve the databases from becoming corrupted.

[mph999]

Revaroo, as always, gives a good answer,

It's preety impossible to cover such a question - experience gives you an idea, but even then I sometimes think I know the logs required, then find out that I need another additional one.

NBU 7.6 has a feature that will tell you what logs are required for a particular issue so things should be easier with that.

TBH, concern yourself with learning how to get logs, let support woory about which logs ...

As you gain experience, you will get an idea about which logs to get, for which type of problem.

The issue is you can't really say , this type of probelm needs these logs, because, depending on the cause, and at which stage of the process it fails, will define the logs required.

For example, a 'tape' problem could be bptm, or robots, or tpcommand, but then again, once you find out more details and discover the tape problem is actually the fact that the bptm is not receiving the resources, it could be nbjm at fault, or, diving a bit deeper, it could be nbrb, or even emm.

You see the issue ...

M

[Marianne]

@Netbackup_user

You will agree that your query in your opening post was answered. New question should be asked in a new discussion.

Please select the post that answered your original question and Mark as Solution.

[Vickie]

Thanks Revaro for sharing the links.....

Thanks to all for their suggestion...