ENLIGHTENMENT - we are using 8.2-fpm (as shown by phpinfo.php in the nextcloud root directory) so we need to edit /etc/php/8.2/fpm/php.ini and then we need to run systemctl restart php8.2-fpm to have new values recognized!
One extra redis question - usually with nextcloud I set it up to use a unix socket rather than listen on the tcp port (nextcloud and redis on the same server). debian 12 already installs redis for rspamd. any reason to leave it talking on port 6379? and if not, I assume we have to edit rspamd - anything else: or just leave it on port 6379?
That#s not ISPConfig specific, its just BIND. As you can see in BIND docs, you must set also-notify only in cases where the secondary DNS server is a hidden server, which means that there is no NS record for the secondary DNS in that zone. All servers listed as NS records are notified automatically by BIND on changes and therefore they do not need to be listed in also-notify. But of course, you can list them there and it should not hurt to do so.
so .... if I add ns11 ip to the also-notify list in all the records on NS10 - (whats the mysql command for that again?) - and change all the ip addresses - and resync - then NS11 should also get all the updated info with itself as nameserver and the ip addresses all changed on ns11 to point to ns11 (migtool completed successfully. - what again was the resync mail-only command to get straggling emails?) that means that I only have to update ns10 and ns11 would be automatically updated as well. saveing much time. now I just have to go to all 200 or so registrar domain entries and change primary nameserver from ns10 to ns11. sure wish THAT could be done automagically. but I doubt namecheap or godaddy have any such ability
The resync command of the migration tool will resync all data (not config) that you migrated. So if you migrated mail only, it will be only the mail data. If you migrated websites and mail, it will be websites and mail. DNS is config only, so it will not change. Regarding DNS server name, have you considered naming the new server like the old one, so you don't have to change DNS server name in all zones at godaddy?
I indeed thought of just renaming the new server ns10. but that would obviously break ALL the certs. I would love not to have to change the DNS server at the registrars... what is my sequence? in particular in phpmyadmin I would need to: 1) change all the ips from oldns10ip to newns11ip 2) change all the nameserver entries from NS10 to NS11. <-- dont need to do this if we changing NS11 to 10 3) change the server name from NS11 to NS10. 4) resync to upload all dns records 5) wait a few days for percolation through DNS world 6) redo all certs? how do we force that? 7) is there a number 7? also have mongodb installed - had to add the Ubuntu repo as its not in deb12 yet. nextcloud working just fine. I bought billing module at some point. probably time to install it, no? does the serial expire? will my NS10 serial for ispprotect work if I move it to NS11 then rename?
now for a real catch-22 - installing debian 12 on a dell r720 server. If I boot in bios mode and boot the usb installer in bios mode all goes well until the grub install phase. it says you only have one operating system should be safe to install on sda. I do so. finish. reboot and have unbootable system. I saw a half page of instructions to fix issue and after a half dozen of them we dome to grub-install - and that fails. (instructions say it should not!). When I change to UEFI boot and boot the installer in UEFI mode - all goes well until we get to grub install phase where it says 'failed to install to grub dummy...' Unable to install GRUB in dummy Executing 'grub-install dummy' failed This is a fatal error now folks said to to back to MBR/BIOS mode to avoid this problem. but scroll up. that does not work. all of this is installing deb12 from my usb stick. any thoughts? rocky installed perfectly no issues same server.
I do not use Dell servers, so I can't help you with that hardware-specific issue. if you can't get Debian to run there, try Ubuntu 22.04 instead.
Is sda the boot disk device? Does BIOS have settings protecting writing to disk boot sector? Is there RAID card or otherwise disk setup is complicated? Internet Search Engines found this: https://askubuntu.com/questions/1008049/installing-ubuntu-16-04-lts-server-on-dell-server-r720
thanks taleman I had already found that page and the grub-install fails (as above says). no, BIOS settings dont have any entries. there is RAID but its transparent to system. sda is the only drive in the system (its a 3.6TB Raid 6) scratching head - I'm going to burn a dvd so that BIOS install will be the only option...
the DVD installer behaved the same way. I SUSPECT the debian installer cant handle >2tb drives in bios mode??? the boot (only) drive is 3.6TB. but I'm perplexed as to why UEFI install fails. server is plenty new enough and as I said, rocky 8 install went just fine. Any idea what the cant install grub-install dummy means?
I think the problem is BIOS mode and fdisk, which may have restriction of 2 TB disk maximum. Discussions about this: https://www.tecmint.com/add-disk-larger-than-2tb-to-an-existing-linux/ https://www.cyberciti.biz/tips/fdisk-unable-to-create-partition-greater-2tb.html https://duntuk.com/how-install-new-drive-linux-larger-2tb-proper-alignment https://thelinuxcentre.com/how-to-add-a-new-disk-larger-than-2tb-to-an-existing-linux/
as it turns out some persistance was necessary. I had to delete the old partition structure by creating a new bogus one, writing it back, then rebooting in uefi mode (in setup) and off ventoy disk - then the install went through properly. something about the old rocky disk structure seemed to disturb debian! but its running bookworm now.
You say nextcloud is a vhost in ispconfig, do you mean that nextcloud is installed inside the web folder? If yes, then this will get migrated as it is, nextcloud version that ships with Debian would not matter in that case. and if nextcloud database is a mysql database you created in ISPConfig for that site, then this will get migrated automatically as well.
doing a second migration from centos to debian 12 and I've received several errors: from the log: 2023-11-07 19:15:51 - [ERROR] API call to mail_user_add failed. 2023-11-07 19:15:51 - [ERROR] JSON API REPLY ERROR: Mail domain - firstcareercenter.com - does not exist as primary but this mail domain does exist - what does not as primary mean?? also these: 2023-11-07 19:15:51 - [WARN] storeMailFilter: Could not resolve 7 to new mail user id for INCOMINGFAX. 2023-11-07 19:16:34 - [WARN] Curl exception: cURL error: [28] Resolving timed out after 10519 milliseconds 2023-11-07 19:16:34 - [ERROR] JSON API ERROR in API call (dns_zone_get): NO ACCESS 2023-11-07 19:16:34 - [INFO] Trying again (dns_zone_get) 2023-11-07 19:16:56 - [WARN] Curl exception: cURL error: [28] Resolving timed out after 10520 milliseconds 2023-11-07 19:16:56 - [ERROR] JSON API ERROR in API call (dns_zone_get): NO ACCESS 2023-11-07 19:16:56 - [INFO] Trying again (dns_zone_get) 2023-11-07 19:17:24 - [WARN] Curl exception: cURL error: [28] Resolving timed out after 10519 milliseconds 2023-11-07 19:17:24 - [ERROR] JSON API ERROR in API call (dns_zone_get): NO ACCESS odd. last migration went flawlessly!
Seems as if the target sever (web server) became unreachable during migration so that API requests to it started to fail.
odd. can I use an internal network ip (as both are on 172.0 internal network) instead of public ip? but there is something else going on. from the migration log: Code: 11-08 14:12:38 - [ERROR] API call to mail_domain_add failed. 2023-11-08 14:12:38 - [ERROR] JSON API REPLY ERROR: dkim_selector_error<br />^M - Arguments sent were: array ( 'client_id' => '0', 'params' => array ( 'server_id' => '1', 'domain' => 'firstcareercenter.com', 'active' => 'y', 'sys_perm_user' => 'riud', 'sys_perm_group' => 'ru', 'sys_perm_other' => '', 'dkim' => 'y', 'dkim_selector' => '', 'dkim_private' => '-----BEGIN RSA PRIVATE KEY-----^M MIIEpAIBAAKCAQEAtNWTTB9/sWjrNeo/anFmvNXMYvtxC2euUJrNnsRnMaKWPDPB^M SXWgRttYc2wcXrh2piq5192+2R5H6khk/rRDr749ITmhu0eta+Tev5me/gOjyXCC^M 9PlMptoVPSmoTmD/vCkFQHXCqlb2dKHmaMvImxQYeUtGbhwYF6wIN0NBo83CVyPi^M iSKz+k0O0khHXDvFS+KpkgNuZPCm2IF3GYpB5M/Chvveyt/9X4iYQaWKE8KeDgUi^M SLQGtr43Lt+LckgdaBoma7f0iUP1NS9cHScafBFwk2gDWumSoVBfHr25ngdNd7H3^M JVGZppWzmEmjhnMsFZ7ZYg5j4+iQfxshmZqshQIDAQABAoIBAFdGlPayZSCssIiu^M 67hlzt3omeJT1HVarui8Le0ogV2VSMj5eZoIqThJeOEMFo0onisIh/VFUmzbjn3m^M PsgnHFvfhwVYGVSX/X0LUu5KKFAUVaJNKQcuou4MZx/D0TUy33pimeRIcZFgs0YL^M NXbyk3k1M7Kmv/E4ofAa2URCs5VoUq4NHh457J3mFzi6e4iKuJCmHHWZuXBVjjVp^M ghc6Y0TUGsu5EIQ2S088ydJD8WaWgs8fqltOuc3W1cgRXowItTbKwfp6l4rDIbF+^M t63FmMxdtbcO/g4nxVHpnbqLwF1pQwt87T1zrRqb+qKn4oR9jGg0H45qJg2dtAE7^M XNTNCoECgYEA6f9oHdGaHA6Dz8hpFp6keHSVG1UFNGYlzE3sG8z1yn2fU6wpUdN7^M Bbe1y9sWSZ73uzu/lQIfwYMrUmRjsARUjhyHpIclkQWeXpELR/v0bSHC8ApKN01Z^M XQCP5YKlVs6KWRL/5I/4uqfKnuDbkYmq8d11JXhx8ml5iuUp0Yvq73ECgYEAxdZ2^M 5AMUJwA1Zlm94AXpx+7uRN06c1iO8Mp10/wTymao+1xacKDmWvOP+CE7Wz3exMwR^M cxfasRFyMt7aqQ9TyEcvkLGw4prakLpQY7vz/u+1xonz0JN7K6B4FsUFRxrXpcxw^M ee90JFa6uMjTszE+hBVbMFroTwisGV/eQokt7FUCgYEAhgG7Ariy5TZ6zlgbttSL^M pGA/tDNwk52cL/DAMvuiF6Dxyd1inHXckl2vQ1vo26iqfxOsmA1rWktgWAMeBAO2^M 2NwF4ndzUCZDmCzApSzpwkX/ebCq6BjoAJb90ZaMUbZfDLwQyDibXWoSZAKqPCL3^M PyRwnlDoEMfOz3iGBnd2yEECgYAgf+C8CD1xQrmC/M8SVCZZb8i4i1ckx7GDoN7G^M WiyuHTs5P6yjv+rXmhdWfa8QiBNmZNeg/pQwbltCPBXTm+75gagwruAN+b5Z9CFs^M oN2uTZag7NyesPuU6SGRLAB389LlK7Z2cX200oKOBH+yei546EE4qlZrVVNgP711^M VK+0mQKBgQDEw8pCS3I3C+TJLMVhALRQ9gScBndS9zqvEc2++ZreZdg8LcQ2SCY/^M /UeChPHm0p30NqRIdhgoBGF+FszjUXHi8zXQTxJ6C8fyDApsxYax3HNzG25n5Gno^M mr7+QUD5NciAKaT8yS/hDg1NqD9pf9flLEgcFJCIQN+6W/c3W/l37g==^M -----END RSA PRIVATE KEY-----^M ', 'dkim_public' => '', '_ID_' => '2', '_migrate_status' => NULL, '_migrate_status_type' => NULL, '_migrate_status_dep' => NULL, '_primary_id' => '2', ), any useful information here?
yes. Yes, the dkim selector field of this domain is empty, but it may not be empty. Older ISPConfig versions had an error that allowed it to empty this field (it contains the word 'defaul' when you create a new domain. Later ISPConfig versions had this fixed and threw an error when you try to insert or save an email domain with an empty dkim selector field. And that's the error that you get, the field is empty and the new ISPConfig server complains about this. You have tow options, one is to alter ISPConfig code on the new system to allow an empty dkim selector, or the better one is that you fix it on the old system and enter a dkim selector.