Holiday Wish List for NetWorker

Yes. This thread is very helpful. We appreciate all of the feedback. We're not supposed to talk specifics on the roadmap in such a public forum, But I thought I'd take a moment to make a few comments about some of the suggestions raised in this thread.

@ Sunman re: LLM - We're looking at LLM, but most of the work that we've done lately in this area has been on licensing itself, as opposed to license management. We generally do some license simplification in every release, and have consolidated a bunch of licenses over the course of the last few years. That said we still have more work to do here. You can expect us to continue to eliminate and consolidate licenses at least throughout 2010. We're also evaluating going nuclear on license simplification, but we're not ready to talk specifics or timelines yet on that.

@ Sunman re: jbconfig - Are there some specific tasks that you think would most benifit from an NMC implementation? This seems like an area where we could take a hard look at an RFE or two and make a big difference in usability.

@ Jason R re: HA - Very, very good suggestion. Thanks for this. As more of our customers move to a B2D + replication architecture this sort of requirement becomes even more important. I will take this back to the team and keep you posted.

@ brerr re: Cloning - We're actively working on improving this area of the product. Not ready to talk timelines yet but stay tuned.

@ Hakan re: Checkpointing - Another area that we are actively working on. We should have news on this capability sooner rather than later.

Thanks again guys for all of your feedback. If anyone would like to get a more detailed view into the NetWorker roadmap, send me a private message and we can work on setting something up.

SnapImage for Linux, Solaris 10 and x64 Windows (including Windows 2008).

I am new to Networker, version 7.6 on Windows and am working in a totally Windows backup environment.

In the Administrator I would like to see the Group Details window in Monitor display the elapsed time in HH:MM:SS instead of just seconds.  I find it hard to divide, say, 3458 seconds into hours and minutes in my head.  Did that not occur to anyone?

I would also like to see more meaningful error messages instead of ones like "Client is not configured correctly".  It must have failed on something, why not let us know?  While you are at it, how about a good log viewer in the Administrator?

I find the Clients wizard unreliable, esp. when adding an SQL backup client.  Sometimes it detects the SQL component, sometimes not.  I can add the parameters manually and then modify it with the client and everything works OK.

Finally, I also cannot see anywhere how the system is writing the tape labels using the lable templates in the Pools section.  Once a tape is in use, that field does not seem to be visible in any of the screens and is not a selectable field in the choose table columns list.  Is also does not show on reports that I can find.  Lable templates seem pretty handy but there is not any place to see what they actually are.  Perhaps it should be in the Volumes screen, since that would be where I would expect to get info on the latest tape volume.  The best I could find was the "Next" column on the Lable Templates screen.  If I wanted to look up a volume by the label template, there seems to be no place to do so.  I would love to have a report showing the tape barcode and the corresponding template lable that was created for that tape.

This would be an holiday list for, say, Spring Break.

Thanks for the chance to make suggestions.

P.S.  If not too late, make MS SQL Server the index storage area.  Storing the indexes in the normal file system seems just too prone to error, even if it works now.

Message was edited by: davek

This is excellent and honest feedback ... truly appreciate it.  It's good to have a fresh perspective for things we may have overlooked (such as displaying the raw data stored by NetWorker).  More troubling to me, are your comments about the configuration Wizard.  Have you attempted opening a case with Support on this?

I'd also like to see if I can get someone connected with you, to discuss your suggestions about label templates, etc.

-Michael

Hi, thank you for the fast reply.  I have not opened a ticket because I am still getting my feet wet in Networker and am getting additional training in house from co-workers who have extensive experience in it.  I have worked with a couple of other popular Windows based products and it is taking me some time to get used to the new concepts or the new ways the old concepts are addressed in Networker.

Long term stuff:

integrate with your CMDB and store backup policies so that when a machine appears in teh CMDB there is a default backup policy and it gets implemented (including installation, etc). You can do it, I had a prototype running in 1995 against MS' SMS.

be able to set break points so that a client becomes a Dedicated Storage Node (e.g., on a 1Gb network, at 450GB of data become a DSN). 1st step is to sent a warning.

Buy IndexEngines. Go on, you know you want to. You could then do a context based restore.... wow!

Integration with Source One... see context based restore. I want one place to manage my archiving and backups.

PowerSnap support for EVA and XP 9000 (you guys may not like it, but HP are a major player in Storage)

PowerSnap support for HDS

Support for more functions on Netapp (See comment about HP)

Fix Cumulative Fulls

Fix LLM so you can add a storage node and it works every time

Do a drive status before you load a tape so that if the drive is unavailable it doesn't then make the label when it returns it to the jukebox... or at least keep the label if it was the same as the barcode.

mySQL module for NetWorker

proper integration of the OpenVMS code with NetWorker so we can get more frequent updates

Integration of NetWorker with Apple TimeMachine, much like you integrate with Micrsoft's backup

Support for Storage Node Client and Oracle on zLinux (I have 3 sites with about 120 clients that want that)

Support for IBM iSeries (There was a texan company willing to do that for you)

Ability to receive streams from other backups so other solutions can integrate and all you do is accept and store the data

Complete the original design for ADV_FT

Give me a job! :-)

I meant Consolidated Fulls.

Here's a short list.

1. nsrmmd needs some attention. Too often tapes (both virtual and physical) end up "unlabeled" because an nsrmmd process either hungs or had a monentary problem. We're a Solaris shop and 50% of the time we'll have to run `fuser -u /dev/rmt/xcbn` to find the offending process ID and then issue it a kill.

2. nsrjobd. It sure would be nice if we could list what jobs are queued and have the ability to terminate a job.

3. Give us back the following: nsrjb -Hvf /dev/rmt/xcbn. If a single tape drive goes haywire we used to use this command in version 6.x to straighten it out.

4. Cloning. Fix NMC so we can see the source volume, source drive, target volume and target drive.

5. Cloning. Stop marking a saveset as cloned at the start of a clone job. Wait until it actually finishes before increasing the number of copies to 2.

6. Give us the ability to list more than one Storage Node in an autochangers "Read Hostname" setting. This would allow a second Storage Node to clone data if the first Storage Node didn't have any devices available.

I could go on but I'll stop here.

re 2. You can do that. You just need to select the columns you want to see.

Load balancing between Storage Nodes for a client. e.g., client starts. Instead of trying the 1st storage node, it looks at all the SN's in its list and says "This one isn't busy, I'll use that". Having said that, to be really smart you may want to tie that in with latency... as in a Storage Node may be free, but the latency to talk to it is awful, so wouldn't want to use it anyway.

You know that Location field on volumes? Use that! You could choose any storage node that has acces sto the pool you need in the same location! I could do that on OpenVMS in 1988!

It has always bothered me that if you are configuring an Oracle RMAN client or NMM application client, you have to specify a username on that client as an administrator in the whole NetWorker datazone!! As I understand, this is because RMAN clients need to be able to do some media management tasks on the NetWorker server.

This means, that a rouge admin on one of the say 500 clients being protected by in the NetWorker datazone, that happens to be a RMAN client or NMM client, that admin, can run nsradmin against the NetWorker server remotely from his client and do what ever he wants.

If I was a customer of a NetWorker datazone backing up my very sensitive production data, I wouldn't feel very good about it.

I hope EMC is working on changing this "feature", that I consider a security hole.

Johannes

Oh, more granular policies for groups/clients.

Why is it that a group policy overrides a client? Shouldn't it be the other way around?

e.g., I want all the clients in this group to start at the same time and have the same policies, except Fred, which I want to be slightly different.

Currently you either have to put fred in a seperate group, or set the policies on all the clients seperately. argh!

For maximum configurability it would be nice if you could at least have the option on a per client basis of saying that the client's schedule overrides the schedule or level set in the group.

That way you could have groups that contain a mix of clients where some have their schedules set by the group, and others have their schedules set at the client level. Currently you're forced to keep these two types of clients separate.

If you do not mind me asking, what kind of policy at group level you would like to override from client? Group does not need to override anything from client level, so I do not understand your request.

BR

Boris

It would be relatively easy to achieve too --- just add a checkbox option to each client resource of "Client overrides group".