Upgrade experiences going from 9.6x to 9.8x

I went from 9.65 to 9.81 this week. The only major issue I've found was imaging computers with recovery partitions not booting properly after imaging. Workaround in this thread

This adds to the list of one more thing for me to test more thoroughly in test environment next time.

@CasperSally did you have any problems with computers not running policies properly after the upgrade? If so, did you have to re-enroll them to fix the issue?

nope @mpermann - not that I've noticed.

I was worried about the issue in 9.8 where clients lose connectivity to JSS, but I'm told this was for the most part corrected if you were upgrading from 9.7x or 9.6x to 9.81 except for one report of issues from one customer with very large install base. I have 1300 computers yet to check in, but too early to see if they just haven't been on network last 2 days or if they actually lost connection.

Side note, I wouldn't do any JSS upgrade without something like CasperCheck in place from @rtrouton unless JAMF builds some self healing into their binary in the future. It wouldn't have helped 9.8 issue, but I've had clients lose connectivity in upgrades in the past that it would've helped.

Thanks for the response @CasperSally. I may wait a few more days then ask you again. :-) I love the work @rtrouton does and will probably look at implementing his CasperCheck solution. I guess I just don't feel like we should have to implement such additional stuff to keep Casper working. This is something JAMF really should be building into their product. If someone knows of a feature request for this that can be upvoted let me know and I'll do that.

@mpermann I wish jamfnation had an area I could easily see things I've voted up recently. My profile - feature requests looks all over the place.

Someone on slack reminded me of this one, worth votiing up

More robust check-in daemon

I upgraded from 9.65 to 9.81 this week also. Managed pool is about 1500 computers. So far, no major issues. I have just two testers running El Cap right now and I've had to update a few of my scripts to define the new PATH location (for jamf binaries), but otherwise, it's been a pretty clean upgrade.

Bonus (for me): something in my JSS instance was causing blank message windows when I used jamfhelper to display a message to my end-users. When invoked from a script, the message would display properly. Since upgrading to 9.81, jamfhelper has returned to its previous usefullness and is displaying my messages properly when invoked via the "Files and Processes" section of a policy.

FWIW we have been cleaning up a mess left behind by someone's hacked QuickAdd. I would vet any ideas of going down that road very carefully.

Those feature requests for health check look great, just wouldn't tamper with the QuickAdd.

@damienbarrett we're seeing the same jamfhelper issue in 9.73 that you were having in 9.65. I look forward to that problem going away as well.

I went from 9.63 -> 9.81 - no major issues, with a fleet of about 300. I did have some issues where users were trying to run a policy in Self Service and getting a "This computer is not managed" error. A reboot fixes that.

I do have one client which has stopped running policies all together (even though it's scoped, and shows as pending in the JSS, the client doesn't pick up on it). I haven't had a chance to look too deeply into it yet.

@damienbarrett @mpermann We were having the blank pop up window issue too and resolved it after support suggested changing the double quote items in our command to single quotes. Worked fine after that.

EDITED to give the right info on double to single quotes.

We upgraded from 9.72 > 9.81 this afternoon. All went well with the installer portion of the update (an improvement over our past experiences) but I'm having some stability issues. The JSS/Tomcat has hung up twice and requires a reboot to get it going again. Working with support on this more tomorrow. We're running on Windows Server 2008 FWIW.

We're in a cluster and have seen tomcat hang on 2 of the instances as well. Clustering makes it complicated as I don't know it's down unless I manually browser to the instances. Fun times.

Tomcat wouldn't start on our 2008 R2 lab server after the upgrade. Found others with this issue, checked all the same stuff (JAVA/JRE in path, Tomcat settings for JVM location, etc) - no joy. Ended up just forcing the installer to lay down everything again with msiexec -fa, and all was gravy. Figured Tomcat just didn't lay down correctly in the upgrade, but can't rule out config issues on my part. This was an inherited lab environment.

Upgraded from 9.65 to 9.81 about two weeks ago. Upgrade on the internal JSS (hosting the DB currently) went well enough, but the JSS in the DMZ ate it and had to be completely yanked off the server and reinstalled, after which it was fine. Out of our client pool of about 600 currently, most machines upgraded their agent just fine but we did have a few agents that broke or had issues and either need a forced inventory update or re-enrollment via quickadd to solve the issue. The total I'm estimating currently is around 30-40. YMMV though as our DB server also had an issue due to failed VM cluster a few days later that impacted machines checking in correctly.