I'm to figure out what has really changed because of some major issues we've encountered since the update.
... such as ...?
Giving context / examples would certainly help giving a better / more accurate answer.
Also - what version did you upgrade from? 9.6 flat? That tends to help put things into context too .
Sorry for the late response. I was looking for more documentation about whats new in EPS, in HIPS, in AV, in software distribution, in replication, etc.
We upgraded from LDMS 9.6 SP2 and had a multitude of problems. Had to re-create all agents again. Problems with AV updating, etc.
All the issues have been worked out with support. But knowing more before going into the update might have helped a little.
I was just ready to install SP3 (from SP2)....
@Support, Can you confirm these issues? I haven't updated to 2016 because I know things break, is it the same for SP3?
We just upgraded to SP3 and had no issues. We have about 9k nodes... We are still in the process of upgrading nodes due to size but we are 25% complete on clients stations with no issues. If anything improvements with CBA and such.
carlos - Your question / assumption isn't overly helpful to formulate a response to (I'm not support per se, but I'll try to be helpful here & put something together).
"I don't upgrade because I know it breaks things" - this is what upgrade / validation testing is for, really.
*EVERY* environment is different and just because "we" have / haven't run into certain issues doesn't mean that you will have an identical experience. The setup of your network security / your used software stack & so on ... will all be different ... may (/will?) affect things differently.
There isn't a straight up "we will BREAK this thing X" that I'm aware of in LD 2016 (assuming you're NOT referring to ancient client OS support. 2016.3 doesn't support XP or Windows 2003 for instance).
As for other things - that's down to...
... what you use.
... how your network is set up.
... and so on.
Which brings us back to validation / upgrade testing -- I'm assuming you've got a dev environment of some sort. Ideally, it should be representational of your live environment.
Run an upgrade there ... see what does / doesn't break (9.6 to 2016 shouldn't be much of a shift, as you're not having to re-do packages / agent behaviors in a way you'd have to do if you're upgrading from 9.0 / 9.5 for instance). Much less a paradigm shift there & more a logical extension on what's been moved towards with 9.6.
It's a somewhat "unfair" question, since there's no proper answer. The fact that I've upgraded several 10,000 nodes to 2016 without much of an issue doesn't mean squat for you, since I am not familiar with your environment. The thing I *CAN* say is this though:
People / companies who invest into proper upgrade testing, do NOT regret it, because they'll identify issues early (in 95% of cases) and can get them submitted / fixed / re-configured / etc before things go into live.
Companies that "test in live" as it were, tend to be repeated disaster houses for reasons that are pretty obvious.
This is usually the fault of overly tight budget strings (some people don't understand how a simple VMWare license can save 100,000 $$-s of lost time as a dedicated test environment) from the higher-ups, which is why I'm *ALWAYS* happy to explain / share some of the real-world horror stories I've experienced in my career that would've been entirely preventable through dedicated (and disconnected!) test labs / upgrade testing.
This isn't a "LANDesk thing" either. If you use ANY management software, you should be doing those things. Anything that's got as much "power" / potential impact on your environment should be treated with care. As a colleague of mine puts it - "We just give you a lot of rope. Whether you pull yourself out of your hole or go hang yourself with it is up to you."
Does that make sense?
It was just a simple question. Thanks.
I was just trying to explain that "it breaks stuff" isn't really something that either can be argued properly, let alone proven/disproven.
It's a more complicated matter than that. That's what I've been trying to get across.
I've done several upgrades to 2016 and haven't had issues. That still won't stop me from suggesting full upgrade testing in any other location / place.
I hope I've not come across as trying to bite a chunk out of you. By no means my intention. It's a simple question - yes - but it's not the right one. That was what I was trying to point out .
To get back on track with the original question - where there's big changes, we're now trying to do / release component specific documents.
So the latest update for 2016.3 for instance has this (various things linked to inside here):
... with the intended service pack update path (don't know if you've attended a LANDesk event recently, but they've been throwing the message around), the intention is to essentially have a quarterly release cycle (so "2016.3 because of Q3 release). And there, 1/2 of the quarters will be "just bugfixes" type updates, while the other 1/2 (such as 2016.1 and 2016.3 were) are "some bugfixes, but a lot of functionality additions / improvements" and so on.
So you should have a better way to expect changes going forward, whereas in the past, it could be a bit "random" that "SP 1 and 2 did nothing but fix bugs, but SP3 introduces some massive change for you" situations).
Does that help you, GJHorn ? We're trying to improve these things as we're going ahead.