Dovecot and Rspamd not working after server crash

J

jstrauss

Verified User
Joined
Jun 1, 2023
Messages
17
Location
Austria
Hello, since a crash 2 weeks ago, my mail-services don't work as expected like before.
My server crashed due a unknown issue and the system was read-only and i had to repair the bootloader and so on. After that my Mail-services like Dovecot, Roundcube and Rspamd didn't work like before and had several issues.
I didn't received any new mails in my mail-client and i couldn't log in to roundcube saying "No connection to storage server". Also my Rspamd Dashboard (https://host.jstrauss.at:2222/CMD_RSPAMD_SOCK/) which worked before perfectly shows a white blank page.

I tried to rebuild all my services and then i received all new emails at once and could log in to roundcube (which took a long time to load), RSPAMD Dashboard didn't work, but after around 20mins it stopped working again. I searched all logs and so on, but i didn't found any useful log information for an error or something else.
Now i have a temporary fix to get a daily update of emails: I have to restart dovecot... So it seems that dovecot and Rspamd are the services which lead to this whole errors.

I made some backups of the full-system to my external backup-server but apparently my Hetzner-Dedicated Server (Backup-Server) did have a Memory-Issue which lead to corrupt backups.... So I have to repair it without loosing any data.

My System:
OS: Ubuntu 22.04.03
Memory: 20GB ECC (Proxmox VM)
Cores: 8
DA-Version: 1.652 (09894e5e75bd47a69827dd5e5c24de5b24759b16)
DA-Build for OS: linux_amd64
Detected Server-OS: debian12_amd64
Channel: stable


Code: 
uname -a
Linux host 5.15.0-83-generic #92-Ubuntu SMP Mon Aug 14 09:30:42 UTC 2023 x86_64 x86_64 x86_64 GNU/Linux

Thanks for any help. Please let me know which logs could be useful for error-debugging/finding.
 
jstrauss said:
Now i have a temporary fix to get a daily update of emails: I have to restart dovecot
Click to expand...

If everything works fine when you do this, I'd say, do that one last time, and then work on making full external backups, reinstalling your OS from scratch, then restoring the backups.

You might well be able to get things to a usable state with rebuilding services and such, but what happens if you miss something and it causes headaches later down the line? You'd have probably wished you did that reinstall sooner.

If storage of the backups is an issue, you could pick up a cloud service VPS from somewhere like Vultr, run it for as long as needed to complete the process, and then cancel it - no need to pay for a full month upfront like you do with some other services.
 
jstrauss said:
My server crashed due a unknown issue and the system was read-only and i had to repair the bootloader and so on.
Click to expand...
How did you repair. Did you also check if the disk was still fine? Because this sounds to me like a disk issue.

I would do as @EthernetServers say to make the backup, check your disks and install from scratch.
 
I repaired the disk with fdisk because something was corrupted. After this my bootloader didn't load and i recovered only the bootloader (/boot folder) from an old backup. After this everything worked correctly except the mail.
Ok i think reinstalling from scratch and applying the DirectAdmin backup would be the smartest way to avoid future mistakes. Thanks.

Another question, which has nothing to do with this, but when i am reinstalling this all. Do you think it's difficult to migrate to LiteSpeed instead of Apache? (Including all http_conf rewrites). Because LiteSpeed faster than Apache or Nginx?
 
I don't think it will be difficult as the conversion can ben done via Custombuild if I'm not mistaken.
 
Ok I created a new VM on Proxmox and reinstalled DirectAdmin on it. Everything now working well again. And it was the right solution, else I would have got problems again in some months/years...
 
It got broken again....
I recently installed new NVMes on my Server and so I also installed a new Ubuntu Version on it and restored the directadmin backup and everything seemed fine. But after some days I'm getting again issues with the mail server, Roundcube takes way too long to login and on mail client sometimes I can't connect and also the Rspamd Webinterface (and the Webinterface Link in the Dashboard vanished).

I noticed something weird, in the Service Monitor dovecot uses up to 13-14GB of memory? Is this normal? I checked the logs but didn't noticed something awkward.
I also enabled the firewall (yes I had it disabled) because I thought some bots are overloading dovecot... But now it is blocking everything and it's still slow and sometimes doesn't work at all.

I have no clue what's going on there... If someone has any idea to fix this it would be very great!
 
I will try to move to AlmaLinux due I read that RHEL should be more stable. I hope it will work there :D
 
You must log in or register to reply here.
Back
Top