recovering from disaster....

Discussion in 'ISPConfig 3 Priority Support' started by craig baker, Oct 12, 2025.

  1. craig baker

    craig baker Member HowtoForge Supporter

    i'm working on repairing some of the listed errors in the mariadb log, but when I go to the table it says has invalid info - it says 'in use'.
    also in the log i'm seeing lines like:
    2025-10-20 9:23:19 341856 [Warning] Aborted connection 341856 to db: 'dbispconfig' user: 'ispconfig' host: 'localhost' (Got an error writing communication packets)
    ouch? whats going on??
     
  2. craig baker

    craig baker Member HowtoForge Supporter

    2 of the tables that mariadb.log reports errors in phpmyadmin has both of them (but no other table) marked as 'in use'.
    how do I repair it in this case?
     
  3. till

    till Super Moderator Staff Member ISPConfig Developer

    Have you tried using the repair function in phpmyadmin?
     
  4. craig baker

    craig baker Member HowtoForge Supporter

    yes it fails:
    c0theeldestgeekx.wp_jet_reviews repair Error Incorrect information in file: './c0theeldestgeekx...
    c0theeldestgeekx.wp_jet_reviews repair error Corrupt
    c0theeldestgeekx.wp_jet_review_comments repair Error Incorrect information in file: './c0theeldestgeekx...
    c0theeldestgeekx.wp_jet_review_comments repair error Corrupt

    but with the one database (mentioned before) excluded, migration completes - with only the c0theeldestgeekx database reporting:
    mysqldump: Got error: 1033: "Incorrect information in file: './c0theeldestgeekx/wp_jet_review_comments.frm'" when using LOCK TABLES [4163/4163]
    =============
    Migration tool run completed.

    so... now if I try to fix or delete the base and do a ./migrate --syncjobs we will be good? hopefully :)
    (or is it sync-jobs?)
     
  5. craig baker

    craig baker Member HowtoForge Supporter

    one more thing I noticed. you aware the auto-installer for debian does NOT create the
    /var/log/mysql
    directory with ownership mysql:mysql and 2750 permissions, so that mariadb cannot write a logfile till this is done?
     
  6. craig baker

    craig baker Member HowtoForge Supporter

    another migration related question - if I implement the ssh-authenticator 2fa on the target server, migration cant handle that I assume? have to disable it at least temporarily? or is there a way to satisfy it?
     
  7. till

    till Super Moderator Staff Member ISPConfig Developer

    I'm not using 2FA for SSH, so I can not tell you whether you can migrate this or how.
     
  8. craig baker

    craig baker Member HowtoForge Supporter

    ok one more issue. multifaceted!
    the datarecovery lab has recovered my 60tb raid 6 array and i'm getting back a 12tb single drive with what looks like the correct directories for debian 12.
    so I see the /var/www /var/vmail/ /var/lib (including mysql).
    now I can attach this drive as an external drive.
    what procedure do I use to 'restore' the emails databases etc?
    the users have /var/vmail now. can i merge the /var/vmail/usernanme folder off the recovered drive into the folder on this server? and rebuild the index? or no need. any rammifications?
    to bring up a website (that was on the new server but not the old one).
    can I create a blaink website, then copy over the /var/www/domain/web folder, and copy the corresponding database from /var/lib/mysql on the external drive to /var/lib/mysql on my working server?
    or do I need to create the website & database in ispconfig, and then copy the /var/www from the external, and copy the database with cp from the external drive ontop of its /var/lib/mysql name?
    what will I run into? I assume I need to fix permissions as it will all be root:root.
    what a disaster :)
    oh - old server is running centos8 dead server was debian12.
    or - alternate idea - can I put the 12tb drive into a server and boot from it? any risks? the data wants to end up on another raid array, but can I build up a raid array on another server, install deb12, and then masscopy from the external 12tb drive? which folders and what do I need to be aware of>

    what a nightmare. but $8k recovery was actually pretty dam reasonable. they are saving my life.
    Help me till!
     
  9. till

    till Super Moderator Staff Member ISPConfig Developer

    Have you considered asking @Th0m to do this for you? What to do depends in the end of what works and what data is there, that#s nothing that I can explain in a few posts here. If you have a new empty ISPConfig system and a backup of the ISPConfig database of the old system and /etc/of the old system, then you can try to restore the ISPconfig database, copy back the website users and groups (only them, not all users and groups), use Tools > resync in ISPVonfig to recreate the basic skeleton for sites and email and then restore data.
     
    ahrasis likes this.
  10. craig baker

    craig baker Member HowtoForge Supporter

    thats a good idea. I will ask him!
     
  11. craig baker

    craig baker Member HowtoForge Supporter

    something I just noticed. from migration from centos to deb12, mailman is NOT supported. it gives a couple of errors regarding that.
    but on the target server, mail does NOT work. and errors say no /etc/mailman/alias file (or something very close).
    why would the mail system on the target system need it?
    emails sent out get stuck in the queue and mailq says:
    8E504FE08B3 3778 Tue Oct 28 12:35:16 MAILER-DAEMON
    (alias database unavailable)
    and when I run newalias I get:
    root@ns11:/etc# newaliases
    postalias: fatal: open /var/lib/mailman/data/aliases: No such file or directory

    those are infact symlinked to /etc/mailman folder that does NOT exist!
    thoughts?
     
  12. ahrasis

    ahrasis Well-Known Member HowtoForge Supporter

    I think because old system have it, so it was copied, but you could try to remove it from alias and see how it goes. Just refer to the latest manual install there is if you are unsure on how and where to remove it. It should be easy.
     
  13. craig baker

    craig baker Member HowtoForge Supporter

    Till looks like I need to delete all reference to mailman before doing the migration
    But shouldn't migration script fix target that does NOT support mailman?
    How to delete it from source server?
     
  14. ahrasis

    ahrasis Well-Known Member HowtoForge Supporter

    If you have done the migration, I think it is easier if you just clean up the new system from its related configs, as the new system, normally setup using ISPConfig auto installer doesn't install mailman.

    About migration tool, as a disclaimer I have never used it before so I would not know, but I migrated some manually and update some from with mailman to no mailman, so I could say it is easy.

    For mailman removal, try: https://www.howtoforge.com/update-t...rom-debian-10-to-debian-11/#-clean-up-mailman

    Also, may be the best if you could check the migration tool help function, if there exists any for mailman exclusion, or just wait for @till to explain and help.
     
    Last edited: Oct 29, 2025
  15. till

    till Super Moderator Staff Member ISPConfig Developer

    That's not needed.

    The migration tool does not alter the overall configuration of the target. So if the target has no mailman, the Migration tool will not insert any config for it in postfix.

    You should not delete it from source server.

    I guess that you copied over config files from /etc from source to target and have not just been using the migration tool. The best fix now is to follow mailman removal steps from Debian upgrade guide:

    https://www.howtoforge.com/update-the-ispconfig-perfect-server-from-debian-10-to-debian-11/
     
    ahrasis likes this.
  16. till

    till Super Moderator Staff Member ISPConfig Developer

    The Migration tool has not copied over the Mailman references, so it's not related to all this. The Migration Tool is not altering the system base installation at all; the system base installation is defined by what you install while you install ISPConfig, the Migration tool does not add Mailman if you did not install it first. The above situation can only happen when you copy over config files from Postfix manually or he did not use the Migration Tool but ISPCopy, which is a tool to make an exact copy of an existing config by copying all files and not importing things in ISPConfig that the Migration Tool does.
     
    ahrasis likes this.
  17. craig baker

    craig baker Member HowtoForge Supporter

    till, I used ONLY the migration tool. not ISPCopy. and I've reported the errors I was running into.
    I copied over NO folders, and I installed deb12 with the automated installer only! on a blank system.


    there are CLEARLY mailman references on the deb12 system. for example in /etc/postfix a grep of the files produces:

    main.cf~:transport_maps = hash:/var/lib/mailman/data/transport-mailman, proxy:mysql:/etc/postfix/mysql-virtual_transports.cf
    main.cf~3:alias_maps = hash:/etc/aliases, hash:/var/lib/mailman/data/aliases
    main.cf~3:alias_database = hash:/etc/aliases, hash:/var/lib/mailman/data/aliases
    main.cf~3:virtual_alias_maps = hash:/var/lib/mailman/data/virtual-mailman, proxy:mysql:/etc/postfix/mysql-virtual_forwardings.cf, proxy:mysql:/etc/postfix/mysql-virtual_alias_maps.cf, proxy:mysql:/etc/postfix/mysql-virtual_email2email.cf

    I did NOT copy /etc/postfix from the old server I've used no tool but the migration (which had to be used several times as I've reported) - so inquiring minds want to know how did these get here?
     
  18. craig baker

    craig baker Member HowtoForge Supporter

    in case you ask, I did NOT install mailman3 on the new server:
    root@ns11:/etc/postfix# apt list | grep mailman

    WARNING: apt does not have a stable CLI interface. Use with caution in scripts.

    mailman3-doc/oldstable 3.3.8-2~deb12u2 all
    mailman3-full/oldstable 3.3.8-2~deb12u2 all
    mailman3-web/oldstable 0+20200530-2.1 all
    mailman3/oldstable 3.3.8-2~deb12u2 all
    python3-django-mailman3/oldstable 1.3.9-1 all
    python3-mailman-hyperkitty/oldstable 1.2.1-1 all
    python3-mailmanclient/oldstable 3.3.5-1 all
    root@ns11:/etc/postfix# apt list mailman3
    Listing... Done
    mailman3/oldstable 3.3.8-2~deb12u2 all
    root@ns11:/etc/postfix# apt list --installed | grep mailman

    WARNING: apt does not have a stable CLI interface. Use with caution in scripts.

    root@ns11:/etc/postfix# apt list --installed | grep mailman3

    WARNING: apt does not have a stable CLI interface. Use with caution in scripts.

    root@ns11:/etc/postfix#


    mailman3 is available but was never installed!
     
  19. craig baker

    craig baker Member HowtoForge Supporter

    im going to build up another server with deb12, and do the migration to that server - see how it works!
    given that mailman gives ERROR during the migration I assume its moving NOTHING over, right?
     
  20. till

    till Super Moderator Staff Member ISPConfig Developer

    Why do you want to do that? This makes no sense: reinstalling your migrated system. I posted in #35 already what to do to fix your system. And that's simply fixing the mailman references as its also done in the upgrade guide for systems that had mailman before. After you did that, postfix should work normally again.

    Actually, everything was moved over successfully as far as I can see from the info you posted, so why do you conclude that nothing was moved over? The migration tool would have told you if it failed very clearly. You can easily verify that, login to ISPConfig on the new server , are your websites and email domains listed there? if yes, you know they were moved over correctly. Then you can also check if there are files in the web directories and maildirs, id yes, you know they were moved over successfully too.

    And I have no idea why you posted apt output about mailman3, a software that is unrelated and not even supported by ISPConfig as it is incompatible with mailman (alias mailman 2).
     

Share This Page