I have a multiserver system, parts debian stretch, parts debian jessie all latest patches. All servers are on ispconfig 3.1.15p2,at least the update says that, but in monitor two of the servers shows 3.1.13p1 but showing green and all ok. On the other side showing log files they are very old means not actual, servers showing the actual ispconfig version show actual log files. What to do where to look, what to check. Thanks for help or hints Rainer
When doing ISPConfig updates on your multiserver setup, did you update master first or last? Did you let the slave do the update settings on master thing?
Seems as if the affected slave server is not able to connect to the master anymore. run the server.sh script on that slave after enabling debug mode to see if you get a mysql connection error.
yes there are errors, found ip address server will connect does not match ip address for that server in user table on master, will correct this. Thanks for the moment
did correct the target ip address at the corresponding ispcsrv user entry in mysql.user, but still cannot connect. Think the generated password of the ispcsrv user does not match. Is it possible to reset/correct this? Rainer
to clarify parts of error message On master server (after correction of the ip address entry) "SELECT User, Host FROM mysql.user WHERE Host <> 'localhost';" shows beside othhers so I asume there ist a password mismatch, that has to be corrected. mysql -u root -h admin.gerdakloos.de -p works with the mysql root password, so on general the mysql connection is functional Rainer
Run an ispconfig update on the slave server and choose to reconfigure permissions in master database during update.
You could have searched how to force ispconfig update. Try this: https://www.howtoforge.com/community/threads/forcing-ispconfig-upgrade.81756/#post-387447
did an upadate with downloding to /tmp, extracting ... update update runs as normal, at least as I know from other update. Answered yes to "choose to reconfigure permissions in master" but still the same error, did that twice. OK ist is just the secondary name server, nothing else, so it wouldn't be too hard to delete and rebuild it, it's a VM so I even have not to go to the hardware box. But how to be shure the master server will be clean in advance. Otherwise is the is an easier way to fix i would prefer. Rainer
You should check the /etc/hosts files and the server hostname of that server, all servers of that multiserver setup must be listed in the /etc/hosts files on all nodes with their correct hostname. If that's not correct before you run update again, then mysql login must still fail after the update.
Most servers, except the problem server, are at one location behind a firewall, in their hostfiles I have the local private IP addresses and the official internet IP address of the secondary name server The secondary name server is at an other location behind a firewall too, so his host file has its own private address and the official ip addresses of the other servers. That worked for years. Now, some mont ago, the location of the secondary name server changed, so I changed his address official IP address. On the secondary name server site the configuration was keept, only the firewall changed. OK I forgot to change the address on same placed, but now thats all OK and double checked. All the host files are OK. I dis update again and the ran the server.sh, same problem. Rainer
Check the details of the ispcsrv* user of that slave server in the mysql database on the master to ensure that it uses the right IP and hostname. If not, check hosts file on master and on that slave server again. The whole problem is most likely just that the master sees the slave as a different IP than what the slave reported as its own IP during setup/update which then results in mysql rejecting connections from that slave. You can also modify the IP manually on the master in mysql for the ispcsrv* user, but take care to use e.g. the phpmyadmin user editor as the user permissions are set very fine-grained which means the user is spread on many different tables in mysql.mysql database.
on the slave part of the error message "Access denied for user 'ispcsrv10'@'62.54.177.160' " on the master in mysql.user that looks good for me here is the complete message from server.sh Thanks for help and patience Rainer
It all comes down to: the user ispcsrv* (you can look up the exact username in the config.inc.php file on the slave server) must be able to connect with the password that you find in config.inc.php file as well to the master server hostname (as specified in config.inc.php file). So you have to find out why this fails now after you restructured your server network and fix that by e.g. changing hosts files or by editing the ispcsrv* user in the master database.
Think I got the problem. What i did 1. corrected the passwords in mysql.user corresponding config.inc.php of slave 2. could login to master with these credentials 3. run an new update seams ok 4. run server script, database not found error 5. deleted everything corresponding of slave 6. delete slave in system 7. setup new slave debian 10 as described by Tills Perefect Server, just components for name Server and PHPMyAdmin 8. added server via expert install correspondig the ispconfig 3.1 manual 9. see new slave server in monitor on master without any further information 10. added a secondary zone to slave, will not finish after half an hour 11.check user.mysql there new slave is listed with ist local private address So in the old running configuration I had a real firewall at my side doing a real address translation and the slave was seen bey the master with its real non private ip address. This offiziell address was listed under the ispcsrvX entry in mysql.user. Now I just have a FritzBox as firewall where incoming DNS request are rerouted to the name server. But there seams to be a Problem with outgoing mysql request to the master. Do you know a solution or where my thinking is wrong.
You can try to tweak this by using /etc/hosts file of the master, try to set the private IP there for the slave server instead of its public IP. Or you manage the mysql user of that slave manually in future which means you correct the IP or simply duplicate it by using the phpmyadmin user manager. and it's important to use the user manager, don't edit it in mysql.mysql directly as the user is vomposed of many records in many other databases, so manually editing in mysql.user will mess it up while the phpmyadmin user manager is able to edit it correctly in all tables.
run server.sh, saw 'ispcsrv19'@'62.54.177.160' so in install the local ip is entered in request the official ip is send. I think the IP from install ist taken from /etc/hosts but the I have to have the local private IP, as I had it on the old location with the the firewall.
Success, after modifiying IP address on master, nor changeing teh hosts on master with local IP address I did an unistall of ispconfig on slave and remove ispconfig db on slave. Set the IP in hosts on slave to the public official ip and resinstalled ispconfig sucessfull. msyql.user naow hat the public ip for the slave, sow the slave in monitor without any additional information, run server.sh on slave, monitor on master is updated correctly. Thanks again for all the help and patience. Rainer