Hi
I encountered several problems while making the upgrade.
1st the dependancy libbsd was missing. It seems it is required for carbonio-confd.service It is obtained from the epel repository
dnf install libbsd
2nd there were a bunch of services that were not enabled which was easily corrected by using systemctl to enable those services.. However carbonio-configd and carbonio-openldap refused to start properly even after reboot. There seems to have been a conflict in a couple of places. I'm not smart enough to figure out where.
carbonio-openldap refused to start because the address was already in use.
carbonio-configd did start, but then it would seem that something somethinig else was trying to start the service and I was getting repeated messages on the console that the service was already started.
I wrote the following script that seems to handle it OK which I run when I reboot the machine from a cron job. I called the script: carbonio_startup_issues.sh
#!/bin/bash # Eli Wapniarski. - May 24, 2024 # # This small script was written to handle startup problems with carbonio after # upgradeing from 24.3 to 24.5 and then upgrading Rocky Linux 8 to 9 # Hopefully this will become obsolete soon :) # # The script is designed to be started after the server has been rebooted and is triggered by the # following cron job that should be owned by root. Of course when adding the line to root's # crontab do not incluide the comment "#" # # @reboot /usr/local/bin/carbonio_startup_issues.sh # # Also, I found that I needed to wait a bit before the script executes otherwise # carbonio-openldap would not start sleep 300 su -u zextras /opt/zextras/bin/zmconfigdctl stop systemctl restart carbonio-configd killall slapd systemctl restart carbonio-openldap
I hope that this is helpful.
@eliwap Thanks for the report, we added a Known Issue for this problem: https://docs.zextras.com/carbonio-ce/html/release/issues.html#rhel-9-netcat-issue
First off.... Thanks for the upgrade to 24.9.5. The upgrade worked a treat. The startup issues that I mentioned seems to be resolved.
For your consideration.... for the RHEL instruction.... It would be a good idea not to include step 4. Because I'm running Carbonio in VM I was able to take a snapshot and risk it. Running "dnf autoremove" broke Carbonio. There may be naive users who do not have experience with the possible ramifications of dnf autoremove will run it and find themselves with a broken server. More experienced sys admins would know about the autoremove feature and would not need to be informed about it
Step 4 is indeed marked as optional, I am not sure what else could I write... What do you mean by "dnf autoremove broke Carbonio"? You can "break" a system in thousands of ways (I also did find in the past some very creative ways of breaking a system 😆 ), but if you can share some more detail, I will be glad to add some into the docs.
Sorry it took so long to get back to you.
What I mean is that running that command removed a lot of stuff and then the server carbonio is running on failed to start correctly. Many things simply did not work
Since "dnf autoremove" has absolutely no impact on the carbonio installation, in my humble opinion, it should not be included as part of the installation step. If an experienced sys admin wants to do that they would know to do that. An inexperienced sys admin would not understand the risks they would take by running that command and if they have no snapshot their server could be in real trouble.