Server 5.9 Profile Manager don't push profile anymore

https://support.apple.com/en-us/HT202487

Ports used by Profile Manager in macOS Server

2195, 2196 Outgoing TCP Used by Profile Manager 5.8 or earlier to send push notifications

2197 Outgoing TCP Used by Profile Manager 5.9 or later to send push notifications

Apple add TCP Port 2197 to Server 5.9. Are You check out Firewall setting?

https://support.apple.com/en-us/HT202487

from this link https://pixelchrome.org/blog/high-load-on-macos-server/

MAKE A BACKUP!!!

1. Disable Push-Notifications in the Server-App <SERVERNAME> -> Settings -> Notifications

2. Stop the Device-Manager # serveradmin stop devicemgr

3. Remove the Push-Notifications for the Alerts Alerts -> Delivery -> Push Notifications -> Edit Notification -> remove all recipients

4. Force Quit apspd(via Activity Monitor)

5. Move (remove) the plist file # sudo mv /Library/Preferences/ApplePushServiceProvider/com.apple.apspd.plist* /tmp

6. Start the Device-Manager # serveradmin start devicemgr

In my case, the plist file had now a size of around 600 Bytes! CPU went down from 100% to 6%.

After that I’ve enabled the Push Notifications and the Alerts.

Or Stop the Device-Manager

Remove the file (make a backup somewhere and store the file there)

Start Device-Manager

Never test this so dont know if good or not ?

Apple had us do this but, sadly, the solution doesn't last. Take these steps at your own risk. (Catalina 5.9 server latest versions)

1. Turn Profile Manager off in Server app.
2. Also in Server App: go into Profile Manager>Volume Purchase for Apps and Books>Configure
3. Open the Apple Business Manager in Safari
4. Download the current Apps and Books Token
5. Click on Replace Token in Server App
6. Upload the Certificate to the VPP/ABM
7. Disable Volume Purchase by turning the slider off
8. Click Done
9. In the Server App Go back into Configure for Volume Purchase for Apps and Books
10. Upload the token
11. Slide and enable the VPP (back on)
12. Click Done
13. Re-enable Profile Manager

This fixes it temporarily but, for me, the issue came back within 30 mins or so and Push services dropped off again.

If you're in an enterprise environment, please call 866-752-7753 and let them know that you're also suffering - this will help escalate the bug fix internally. Hoping and praying that a patch is released in the next few days.

Hello again diamondbollox,

Yes uppgraden from server 5.7 to 5.8 and 5.9. Also same with Catalina (so all is upgraded)

Before i try remove the serverapp to trash and reinstall it but no change.

My servers is still upp and running (not tuch any tasks and not restart the servers yet)

After i replace the VPP token and disabled the servive "profile manager and VPP" and uppload a new

token to ASM (export from PM)and download new to PM, after about 30 min i have all the active tasks i hade before 200+

I soon try to restart one of the servers to se wat happen.

Maby it works for clean install of Catalina and use PGAdmin or DBeaver, or pg_Dump to restore the DB (not tested)

I write if i come up with something more. :)

Unfortunately (in my case), even after the applying server app update that was embedded in the OS 10.5.3 update, I'm still having the same issue.

My specific issue was/is whenever I make a config change and hit "save", an active task to push out the changes is not being created. Whereas before while using Server 5.8 and Mojave, these tasks would indeed be created after hitting "save" to push a config change.

My workaround (albeit lame) is once I submit config changes, I reboot the server.

Once rebooted, the active task(s) show up.

This process is also necessary if I try to use the update info function, once any config changes are saved.

At this point, I'm just thankful the server rebooting is consistently working whenever I need to push out config changes/updates.

Some updates from today.

Profile Manager and ASM. One of my servers stop today push 2020-02-20

This server run for a couple of weeks (after i do this steps)

So i do the steps again, and the push start as before. :)

If it stop again after some weeks, i do the steps again and again... until Apple release a new server app.

1. Download your VPP Token from ASM (to your Profile Manager server)

2. Disable Profile Manager ((off)

3. In Profile Manager Volyme Purchase for Apps and Books, Configure/Replace Token

4. Disable VPP

5. Start Profile Manager (on)

6. Start VPP ( if you not can start, replace the VPP token again and click (on)

7. In Profile Manager Apple School manager click Configure, Replace Token and export

8. Open ASM Settings your server and Edit, transfer new.. (replacit from export Profile manager)

9. When finished click Ready

10. Now download the Token from ASM to Profile Manager and in Apple School Manager add the new token in choose..

11. Continue and done.

Sorry for my spelling...

AlaskaFromQc,

Similar issues here. My Profile Manager stopped triggering update tasks like it did before, whenever I make configuration change to my clients. I found if I reboot my server, once it comes back up the update task is active and running.

Like you, this issue did not appear until I updated to 10.15.1 and Server 5.9.

Steps. If you not have ASM/BM not tested..

1. Turn Profile Manager off in Server app.

2. Also in Server App: go into Profile Manager>Volume Purchase for Apps and Books>Configure

3. Open Profile Manager admin pages/Apps and Get more apps. Login with your VPP account and download the token

4. Click on Replace Token in Server App

5. Disable Volume Purchase by turning the slider off

6. Click Done

7. Look in Profile Manager, Apps should be empty and look for the active tasks

8. Re-enabled the VPP in serverapp

Apps should be back as before, and Active Tasks (same before stop working) Then leve all as it is.

If the push stop working after upgrade to server 5.9

and you do this step and the push starting working again

the server has very much to do to update apps and settings

so leve all to working...

not restart not touch any tasks.

Reinsatll serverapp i think not help beacuse you have back all the stuff from postgres DB

This what i think have to do with VPP and tasks, the steps above start the push again,(maby) after i upgrade to 5.9

one of my servers i running good for 3, weeks, after 3 weeks, i do sudo command to clean every Active Tasks and Completed tasks

after that push stop working.Then i do all this steps and like i say not touch anything.

Still working.. so far 2 weeks now.

If it still not working try the steps again. :)

Sorry for my bad spelling ...

Hey all,

just a little hint was the problem was on our system. We had exactly the same issues like apsdp eating one cpu core and no Push communication. Tried all things ostersund postet but had no luck.

Problem: Apple has changed a Port for communications with the APNs from 2195 and 2196 to 2197 which was blocked from our firewall.

Solution: open Port 2197 outgoing in your datacenter firewall and kill once again apspd

https://support.apple.com/de-de/HT202487

edit: as i have seen afterwards yu99 has postet the soluion yesterday. I didn´t see that

Screw this, I just found a full system backup of my server running mojave/server 5.8 and it runs perfectly just like i never upgraded !

Will wait few months before doing catalina /server 5.9 upgrade again

Quick question for you @osterlund. Our server 5.9 was originally a 5.7 server which was upgraded to 5.8 and then 5.9. Are your working Catalina 5.9 servers all fresh installs of Catalina and macOS server or were they also upgraded? I'm thinking that there are some residual issues that carry over unless you have a fresh server build. We might try a completely fresh Catalina install and then a migration from our unstable Catalina 5.9 server to that.

I've also taken your advice and requested that my colleagues do not touch Active or Completed Tasks for the time being - things do seem to have settled a little since leaving those alone. Thx again.

@AlaskaFromQc - I think you hit on something, when you mention Apple abandoning PM.

I've seen for sometime now, even Apple folks strongly recommending third party MDMs, like Jamf. When I have explained to them that from a budgetary standpoint PM is all we can afford, so I have forced myself (as I have for years) to learn and leverage strictly Apple utilities, the same Apple folks tell me that Apple really doesn't like to advertise PM as being a "real" MDM, and that the only reason Apple even still provides PM, is to provide a basic MDM template for third parties like Jamf and others to use when developing their own MDMs.

Pretty depressing, in my opinion........

Hello again i have 4 servers (5.9) 2 working great and 2 stop working. whe have about 6000 devices split on this servers.

The most devices is on the 2 stop working. I remember i sometimes clear active tasks and completed task, i do this for this servers for i time ago then it stop working. The other 2 servers i dont do this and this still working. I think it is (what i belive somthing to do for push and tasks) in the logs for postgres it is ((magic...somthing..)) is nil (should be NULL hmm...

Yes but my 2 servers still pushing all working fine, as i said before. I not testing to do this step after server stop working but should working after uppgrade to 10.15.3. stop profile manager and try stop vpp replace token start profile manager,a start vpp

maby uppload token to ASM/ABM and back to server.. maby.

So this servers working for now ... see tomorrow if working.

Thanks again ostersund.

I am indeed running PM in a standalone config.

There is no VPP or ASM account tied to it.

I ended up going the scorched earth route, and completely wiped the server, and restored it from a Time Machine backup that consisted of OS 10.14.6 and Server 5.8.

Everything is working well again.

I'm thinking I'm going to stay with OS 10.14 and Server 5.8 for a while, and perhaps try again with 10.15 and Server 5.9 later this year.

Thanks again for your great troubleshooting ideas.

I can tell by the responses your solutions greatly helped many.

Matt