The Future of NetWorker

Hi Margus,

LV is cool thing, but more suitable for cloud backups.  If you have issues within network infra inside organizations or DC then those should be fixed.  I do not see backup application being responsible for that.  Surely, it is nice to have, but also additional effort, load and risk to code so I doubt engineer will do this.  On the other hand, during EMC world I did came across this idea from one EMC engineer, but it was not related to NW.  As NW relies on TCP segment provided by platform on which it runs, I see more this happening at platform level than application one (if it ever happens).

Key management is usually associated with key managent provider which is dedicated and outside backup application as per best practice (when talking about security, so something like KMS appliance would fit this).  You are most likely referring to somewhat more basic encryption model as NU had it for quite some time (and I guess few others too).  Honestly, I did expect EMC to come up with something like that as they have RSA division under their hat, but they didn't.  I assume reason might be that few years ago research showed those who needed would use certified applicanse and more dedicated solution (which is more costly, but also more efficient of course).  Whether this is the reason or not, I don't know.  NW does supply AES 256 encryption I believe and in such case that should be enough. 

Software based compression exists as per legacy reasons and it happens on clients.  Nowadays, if used I guess it is used by very small number of people.  Compressions and encryption do not go together so if you wish to have client encyption - with it you get compression already.

Pasuing sessions would be great thing to have.  I assume it would need to rely heavily on checkpoints.  I have very busy 24/7 server, but I used nsradmin to pause/stop activities (in my world, I rarely have backup which lasts more than an hour - exceptions being few larger BCV backups - and all updates are planned so planning "downtime" should not be an issue).

There are definately options available.  The DD160 datasheet actually has a section on remote office. :-)  If you would like additional information on options, send me a message off list and I'll have someone give you a call.

Thanks,

Allan

Thank you Allan for offer. We are currently in contact with local EMC and will see if can be found an acceptable solution for us :-).

MargusR wrote: Hi by the way MaRGUus

Sorry, my bad... corrected.

Hi

by the way MaRGUus

Hrvoje,

good points about future features.

Tnks

Margus Raasik

One more thing to add.

Start one group right after another one ended.

Some dependence.

Example if group a finises then this finishing triggers start of group b or c or event both of them.

Wbr

MR

I'd like to see the drive-sharing capabilities of AlphaStor written into the NetWorker server itself.

And I'd like to see quite a few of the suggestions already made. In particular the one about correlating HA services to their current hosts, and the aftd improvements.

As a very large site we run three NetWorker servers to cope with the load. So the debate on "one large NW server vs better communication between multiple NW servers" is highly pertinent. We also have to play the "which server did my client use" game for restores. We had to introduce more servers because we broke the limit for the number of physical devices. However there are pros and cons each way. Multiple NW servers does provide a degree of resiliency. If you have three and one goes down, at least two thirds of your backups ran. On the other hand, its an administrative nightmare.For example, adding a new backup for an HA service requires tha you know on whic server the platforms are configured, so as not to split bits of the whole over the multiple servers. On balance, I'd rather just have the one if possible.

I too would like to see Networker become a more mature product which would have some of its sister product features integrated into it.  Avamar, Data Domain, and Alphastor each offer the end user a rich set of tools.  Why is it that EMC is incapable of intgrating these toolsets into a native Networker?

Or the question could be posed in a different manner.  Avamar is a great product but has the ugliest interface ever.  Why is it that the NMC cannot be used for Avamar?  Why is it that an AFTD is required by the Networker controlling host when it has an Avamar grid included in its strategy?  Why are we not seeing more work being done to support EMC's customers who have Avamar, Networker, and Data Domain along with other EMC products. 

The level of disconnect is not only shameful in the product but it is even worse when it comes to the sales and support folk.  It's as if the DD folk don't want to hear about Networker or Avamar.   What happens if the customer happens to have all three products and they also have Alphastor in the mix?    This becomes kind of a three ring circus as you watch each of the reps say that this piece is not theirs and they can't comment.

The world has changed.  My company uses physical tape libraries where they make sense, Avamar for remote offices, Data Domain appliances to replace aging tape libraries, and Alphastor to help manage the tapes and share tape drives.  (And I won't go into DPA or SourceOne or other products where again the sales force falls short).

I would love to see the NMC evolve and be the 1 stop for controlling all the EMC software applications I currenlty use in the backup environment.  I would love to see Networker integrate more features which would help us manage the Data Domain appliances, the Avamar Appliances, or even Alphastor. 

I would like for the Networker controlled Avamar backups to have the NW metadata written to a DD volume instead of an AFTD volume.

Now having said all that...  I sure am glad I am not working with the same networker version I was working with 10 years ago.  Upgrading from one level to the next is getting easier and easier.  Bravo, job well done.  As for stability, again, things have imporved dramatically. 

Keep on improving the product and make it more integrated with its sister products. This type of innovation will drive the market.

Hello Adanac and all,

I would like to chat with you about our future plans if you like. I am not yet prepared to post anything on the Community just yet but would be more than willing to talk to you about how we plan to address your very valid concerns. Tha tor you can reach out to your EMC Sales rep and have them get the Product Management team on the line to show you the roadmap and what is coming. I thikn you will find it quite compelling.

Cheers,

Steve

One thing I'd like to see is change of calls used by NW (client) so if there is hanging network mnt point I do not get session hanging even though I backup only local file systems.  One may say you get the same if you run df command, but I can also say that you can run df with switch which instruct it to check local fs only and thus does not hang.  By default, NW does backup of local fs only thus that should be taken into account too. With increased NAS storage shares people may just get more and more hit by this. This has been on my wish list for years now, but nothing was changed (despite RFE existing).

Hi Hrvoje,

Do you have the RFE # for this?  I'll check on it..

Thanks,

Allan

Hi Allan,

I do not have it, it was long time ago.  I did quick searh on PL and found esg56045 which seems to fit description and mentions LGTpa72728. However, on second reading it seems to suggest that in that case user was backing up hang NFS share which is not my case.  In my case, let's say, you have

/

/usr

/var

/nsr

/opt

/foo

/bar

/usr/local

/usr/share

All above are local except last one which is NFS mnt.  Saveset All will pick all local by default and that's ok. However, if for any reason NFS is stale you will see save at the end hanging.  As soon as you correct this it will exit with success. This issue has been around since I started to work with NetWorker

And while on subject of enhacements, how cool would it be to apply virtual license model to blades too?

I would like to have either a date range between the oldest and most recent saveset or just the date of the most recent saveset added as a column in the tape volumes page on the media tab.

When I'm looking for tapes to recycle I need to look for the date of the savesets on the tape because I need to leave one set unrecycled per month.  This would save me from having to right-click, view save sets for every tape.

Thank you

Hi Shrimper,

Backupp application does autorecycling management on its own - so ideally you shoul not do that. On the other hand, people may get into the issue with resources and get rid of some savesets. In such case you may wish to delete some tape you do not think is needed, but this is tricky as you have dependencies between full and subsequent incrementals.

It is not clear to me from your post why do recycle tapes on your own if you have enough of those, but perhaps addressing that would be much better choice. Nevertheless, what you are after is easily obtainable from CLI.

Long time ago, I propossed RFE (no, I do not know what the RFE number is) for Legato to implement something I called nsrtree. It is nothing more than MS DOS tree command, but instead of folders it works against backup sets and you can easily filter it by client, saveset or even volume (thus you see dependency tree). Recently I've been giving a thought of scripting this myself, but didn't find enough time (yet).