Issue Dovecot IMAP Mail error - No SSL-Connection over Port 993

[joschli]

Hello, I an error with my dovecot mailserver on Plesk.

Even if I want to load the Emails over the Apple Mail or Outlook Client, I'm getting the following error:
"Mail could not establish an SSL connection to the server "example.tld" via the standard ports. Make sure that this server supports SSL and that your account settings are correct."

All certifications and DNS-Records are correct and I think the IMAP failure comes from the dovecot server. The postfix SMTP connection is working correctly and I am able to send emails over the mail clients.

I think the following error from the maillog is responsible for the failure, but I can't get any Information from google to solve this failure.

May 22 11:21:27 h2830548 dovecot: imap-login: Disconnected: Auth process broken (disconnected before auth was ready, waited 30 secs): user=<>, rip=91.43.240.92, lip=81.169.166.57, session=<p680JjmmW9VbK/Bc>
May 22 11:21:27 h2830548 dovecot: master: Error: service(imap-login): fork() failed: Resource temporarily unavailable (ulimit -u 62987 reached?)
May 22 11:21:27 h2830548 dovecot: master: Error: service(imap-login): command startup failed, throttling for 2.000 secs
May 22 11:21:29 h2830548 dovecot: master: Error: service(imap-login): fork() failed: Resource temporarily unavailable (ulimit -u 62987 reached?)
May 22 11:21:29 h2830548 dovecot: master: Error: service(imap-login): command startup failed, throttling for 4.000 secs
May 22 11:21:33 h2830548 dovecot: master: Error: service(imap-login): fork() failed: Resource temporarily unavailable (ulimit -u 62987 reached?)
May 22 11:21:33 h2830548 dovecot: master: Error: service(imap-login): command startup failed, throttling for 8.000 secs

When I got to the Plesk mailserver settings, change nothing but click on save changes the IMAP emails come in for a few seconds and then the failure returns. This sounds strange to me and I have no idea why this happens.

I hope someone is able to help me.
Best regards
Jonathan

 

[Martin Dias]

Can you provide the output for these commands?

# plesk -v
# grep -v " 0$" /proc/user_beancounters
# ps -aux | grep dovecot | grep -v grep | wc -l
# ulimit -n

 

[joschli]

Can you provide the output for these commands?

# plesk -v
# grep -v " 0$" /proc/user_beancounters
# ps -aux | grep dovecot | grep -v grep | wc -l
# ulimit -n

Thanks for your fast reply. I have attached the Information, you asked for. Best regards.

 

[Martin Dias]

Is your server a VM on Strato? I almost suspect that the issue is with their Host Systems, seeing similar issues in the past days on their servers.
From old tickets it could be an issue on their side on the default setting of the fs.file-max configured on the Host

 

[joschli]

Hello, thanks again for your answer.
Yes, it is an Strato root Server. I also had to reconfigure my whole server a few months ago, because they updated their servers. The system was crashed there.
What do you think can I do to solve this issue? Should I write to the Strato support?

 

[Martin Dias]

Yes, I would recommend contacting them, as on first sight your configuration is fine.
They need to check the HostSystem on their side, you may also point to the "fs.file-max" setting that we know caused issues in the past (although may not be the same this time)

 

[ellog]

i have exactly the same issue with a Strato vServer, they told me this is a standard image it can not be there fault

# plesk -v
Product version: Plesk Obsidian 18.0.34.2
OS version: Ubuntu 18.04 x86_64
Build date: 2021/03/25 10:00
Revision: e5a6bfface98f8934d483a7a2fcdcc53195bda68

# grep -v " 0$" /proc/user_beancounters
Version: 2.5
uid resource held maxheld barrier limit failcnt

# ps -aux | grep dovecot | grep -v grep | wc -l
60

# ulimit -n
1024

# /sbin/sysctl fs.file-max
fs.file-max = 2097152

Max Tasks was at 60 at reached easily 60 after 2-3 Minutes

 

[ciB]

We also have the problem with a Strato root Server. The error occurs when sending mass mails / newsletter to our clients. Is there any way to fix this?

 

[themaxx32000]

@ciB check

systemctl show --property DefaultTasksMax

which is set to 60 on my end. I increased this value to 300 through

/etc/systemd/system.conf

on another system which seems to work fine now.

 

[themaxx32000]

also check

cat /proc/user_beancounters | grep numproc

for the numproc limit on your machine.

DefaultTasksMax must be equal or smaller than numproc limit and you have to be aware that setting DefaultTasksMax == numproc limit basically means:

One process can use all of the available ressources (i.e. tasks allowed) which is presumably not what you want..

Lastly DefaultTasksMax can also be a percentage: If numproc limit == 700 and you set DefaultTasksMax=75% it will result in a DefaultTasksMax of 525.