It's your call, but Grant's comments are really pretty positive, so I'm surprised at your conclusion.
Denny is right - I think there are some great advanced in 9, and I'm happy I upgraded. I'll look forward to seeing the new release that will be completely ASP.NET.
The new Universal Poller works great - I'm still learning my way around the new interface, but it shows a lot of promise to expand in the future. I can finally monitor my Brocade Fiber switches on my EMC SAN thanks to the support for SNMP tables.
There are still some bugs, but nothing that's mission critical or shutting down my operations. My alerts upgraded perfectly (at least, I think they did!) and I had no problems with my custom HTML and web interfaces. Overall, I'm extremely satisfied.
I think bugs like these are things you're going to come across - even with beta testing, you never get everything out until you make it available to all the people using it across so many diverse environments. On top of that, I think Solarwinds has an excellent history of helping to resolve these issues in a timely manner, not to mention help from the Thwack Community. Don't go bashing 9.0 until you've tried it and experienced some issues - I'm happy with the upgrade and plan on working through the minor problems I've had. The benefits GREATLY outweigh the caveats.
OK, my experience was not good but it was that way for a not so obvious reason.
I upgraded from 8.5 to 9.0 today and when it got to the point about the license I forgot that the box didn't have access to the Internet. So the install was cancelled part of the way through. What resulted was an install that looked like it went through with no problems but was about 75% completed.
Things that where broken:
1. Web site would give me an error page that I needed to start the service. When you looked at the services screen it showed that all the services had been started.
2. Universal Device Poller would try to start but would fail saying that it could not connect to the server.
I opened a ticket but after I did that I decided to re-run the installer and tell it to "repair". This has fixed everything.
To all those thinking about waiting for 9.5. Universal Device Poller alone is worth the upgrade in my book.
As I was finishing the post Support called me. So that response time was pretty good also approx. 9 minutes.
Our upgrade from 8.5.x to 9 went smooth, but we see the node lookup by search take 4 times as long now vs version 8.
Just upgraded from 8.5 to V9. Only one issue so far and that was with custom property editor. We had a few fields setup and after upgrade they were gone. When we had these nodes setup by these fields after upgrade they were not.
We went in to readd them and orion stated the fields were allready present. Refreshed the web page and now my missing fileds were back and nodes were sorted by those fields.
Only issue so far.
Our upgrade from 8.5.x to 9 went smooth, but we see the node lookup by search take 4 times as long now vs version 8.
Could you open a ticket and send us diagnostics?
I got the same error why I try to run universal device poller: "an Item with the same key has already been added" then " Cannot contact server. Exiting application"
BBusbey, can you open a support ticket and get some diagnostics to us?
Did this work?:
"Go ahead save your Reports and maps folders in C:\Program Files\solarwinds\Orion. Then you will need to uninstall and re-install the application, this should resolve the other Application issue.
Sean Martinez "
Wanted to see if you got it working before I waist a lot of time to re-install.
The upgrade went smoothly right here. I don't see any major issues yet.
The only thing I also discovered is the x- axis for graphs that is not working correctly. When I want to see a graph of for example 7 days. The x-axis shows hours in stead of day's which would make more sense. This changes in days when I change the sample interval to 2 hours.
Nope didn't work for me,
I have a ticket open and a webex session scheduled for this morning with the development team. I will let you know what the fix was after that if we get it resolved.
I'm a newbie here but I figured I'd contribute to the community and share my upgrade experience. Took a hair over 30 minutes with ~6000 polled objects in my database on a Remote SQL server to upgrade. Not a single error much to my surprise... everything looks fantastic. For those of you who have installed modules and removed them and then suffered the dreaded "The NetFlow module banner is still on the web page" error, Orion v9 installation removes this. It doesn't remove the Netflow service in my experience but it might be better for you.
In any event, v9 is great and I can't wait to give NetFlow V3 another chance.
I may have spoke too soon. The upgrade appeared to go as planned. I did it during the day and had already collected a number of Events and traps earlier in the day so everything looked to be normal. I've checked Orion out the past few days and despite the fact that the SNMP Trap service is shown as being started in MMC it's no longer displaying in my Events view, Traps view, or Network Summary Home view. But I know it is receiving traps, because I would've known if all of my network devices were modified to be sent somewhere else or what-have-you.
FYI, Syslog still works just fine. I can get current data anytime by switching to the Syslog menu bar. Weird. Anyone else seeing something along these lines?
What a horror story. We had 2 hour window of Down Time when we could upgrade from 8.5.1. First off, the customer portal had initailly posted Orion 9.0 SLX. So I DLed it and installed. Tried to license and got an error. I return to the customer portal and notice that the SLX version has been replaced with SL2000. Fine.
I DL Orion 9.0 SL2000, install and license. I think all is well. The server then prompts for a reboot. Upon reboot, the configuration wizard automatically runs. Of course, it replies with an error when trying to access the netperfmon DB. "Login failed for user"." No where in the config wizard can I specify the user ID and PW I want Orion to use. It just goes out and trys user ID ".
So, lucky for me, I still had a copy of Orion 8.5.1 install package. I run that package, and get to the config wizard once again. Same error. So now, after attempting to upgrade to 9.0 we are left with a completly broken Orion package. The configuration manager for both 9.0 and 8.51 error out.
I remember when initially running the config wizard way back with Orion 8, I was prompted to supply SQL credentials. That no longer seems to be the case.
My advice,... stick with 8.5.1, let all these horror stories play out, and wait for 9.5.
When I installed Orion 9 I saw the same thing, but if you run a repair then the wizard, you will see the connection string prompts.
There are a few reg keys you will have to delete to fix this.
i removed all but the default key and was prompted. of course make a backup first :)
Hope that helps.
Thanks for the tip! I had opened a ticket and support seems to really be pushing users to upgrade to 9.0 Usually when opening a ticket, weeks, if not months go by with no progress. I opened this ticket and a tech called me within 5 minutes asking to set up a gotomeeting.
Thanks for the advice. I did not notice anywhere to run a repair. After installing 9, The server reboots, and goes directly into the configuration wizard. Normally, when running the config wizard, you get the option of "repair" and you can choose website, database, etc. After installing 9, I no longer have that option. The config wizard attempts to run, and I get an error.
I think I will wait for a service pack or 2 before attempting the upgrade. Our entire New York State agency relies on Orions monitoring and alerting. Down time is on Orion is critical as I has to beg and plead to attempt this upgrade. After having to revert back to an older snapshot of the server to get Orion back up and running, I dont think the powers that be will allow another upgrade attempt anytime soon.
Orion 8.5.1 SP3
No webex session today I guess they don't know WTF is going on with ver.9 and I am now full hot since I just now received an email stating that they are confused and no word all day until now.
I have 2 tickets open
49810 dealing with the universal poller error saying "an item with the same key has already been added" and it will not open.
49582 problem with virtual interfaces all defaulting to 10 Mb instead of a gig or more and this is happening on my 6500's firewall service modules, F5's, Bluecoat device, basically anything with a virtual interface it wants to make it 10 Mb and this was never an issue in 8.5.x or anything before.
So far I have to give support an E for effort andd that is about it.
I looked into getting updates regarding those cases and the technician for 49810 is out sick today and the 49582 issue is test tracked, which means development is still seeking possible causes/resolutions. We'll see if someone from support can reach out to you today.
I just now received an email stating that they are confused
When we get a bug reported to development, we try to reproduce it. That's a part of the Dev team, and sometimes they don't have all the info they need to reproduce. I think the email was trying to gather more info to work on your bug.
Thanks for the advice. I did not notice anywhere to run a repair. After installing 9, The server reboots, and goes directly into the configuration wizard.
Mike, running a repair is done from the installer or from add/remove programs. It's different than re-running the configuration wizard.
I too ran into the same problem. Instead of changing all of the them to Gig if you like you can add
" InterfaceName not Like '*VLAN*' " under the edit function ( top 10 interfaces by % Utilization) under filter interface. This will take anything with the word VLAN out of the top 10 view. Makes the top 10 still usable until the fix comes out. If your description is coming up with non routed then replace VLAN with what ever key word you like. You can also do more then one key word
example: InterfaceName not Like '*VLAN*' AND InterfaceName not like '*side*' AND InterfaceName not like '*Extranet*'
Here my top 10 interfaces by % Utilization will not have anything with VLAN, side and Extranet included
Down time is on Orion is critical as I has to beg and plead to attempt this upgrade. After having to revert back to an older snapshot of the server to get Orion back up and running, I dont think the powers that be will allow another upgrade attempt anytime soon.
Maybe you could use this experience to build a solid business case justification to get "the powers that be" to drop a dime and get you a test/QA instance of Orion/NTA/APM/Cirrus (smallest license versions.) I know its a "tough sell", but it is well worth the investment.
just a suggestion...
I've been running a pretty basic Orion V9 (upgraded from 8.5.1sp3) for the past week now.
With 8.5.1 sp3 we were finally getting down to only a single open case with Solarwindssupport - and a few issues which was supposed to be fixed in V9.
With V9 however we've had numorous NEW problems - at this point I'm about to reinstall V9 as the "Module Engine" craches whenever someone enteres the node view and support has been unable to find out whats wrong.
I hope Solarwinds will soon realise that fixing bugs is more important than adding new features (and new bugs...)
I did a new installation on a fresh server...install went without error...but since then, I've seen two problems that have me worried...
1. The Service Manager, with the events open...I've notice that on doing multiple refreshes, I've had the Manager lock up...had to kill in Task Manager...this happen a number of times yesterday....
2. Last night we had a number of down alerts...the came back during our reboot period (3 to 5am)...at 5, all the downs were reported...they were not down...I rebooted Orion and they all came back...wondering if a service got in a weird state...if this happens again, I will check for that...
we had a number of down alerts...the came back during our reboot period (3 to 5am)...at 5, all the downs were reported...they were not down...I rebooted Orion and they all came back...wondering if a service got in a weird state...if this happens again, I will check for that...
I can confirm that this is indeed an issue - I've experienced something simular with Alert being triggered after the problem was resovled. Could not find any other explanation for it.