Gateway error

Coffee_is_life

Hello Forum,

this night i got called because the server didnt deliver mails.
after some investigating i found following error in gateway.log:

Tue May 15 00:31:01 2018: [notice ] [25814] Starting worker process for IMAP request
Tue May 15 00:31:01 2018: [error  ] [ 9209] ECChannel::HrEnableTLS(): invalid parameters
Tue May 15 00:31:01 2018: [error  ] [ 9209] Error switching to SSL in STARTTLS
Tue May 15 00:31:01 2018: [notice ] [ 9209] Disconnecting client.

after this time no one got mails, but dagent sends everything to inboxes (no error there)

i’ve found and old ticket, but this wasnt apply here.

gateway config:

# enable/disable POP3, and POP3 listen port
pop3_enable     =       no
pop3_port       =       110

# enable/disable Secure POP3, and Secure POP3 listen port
pop3s_enable    =       no
pop3s_port      =       995

# enable/disable IMAP, and IMAP listen port
imap_enable     =       yes
imap_port       =       143

# enable/disable Secure IMAP, and Secure IMAP listen port
imaps_enable    =       no
imaps_port      =       993

at 04:00 the gateway service has a scheduled restart, after that everything worked again.

coffee_is_life

fbartels

@coffee_is_life said in Gateway error:

after this time no one got mails, but dagent sends everything to inboxes (no error there)

so “no one got mail” actually means users fetching mails with imap did not get new mails in their imap clients?

Coffee_is_life

i’ve got reports about even z-push users didnt get mails. - but there is no gateway involved, i know.

meanwhile im searching for other/more logs to explain this behaviour. z-push.log stopped yesterday at 23:52 and the next message is dated to 02:07 - nothing between.
no errors or suspicious logs in /var/log/messages,

can an I/O timeout lead to this behaviour without logging something? - my backup via veeam runs around this time

coffee_is_life

fbartels

@coffee_is_life said in Gateway error:

can an I/O timeout lead to this behaviour without logging something?

I cannot imagine that. but then again z-push logging started before you restarted the gateway.

were there mails delivered notifications in the dagent.log in the time between 23:52 and 02:07?

Coffee_is_life

the dagent. logs delivery all the time. z-push stopped at 23:53 and started at 02:07 - no errors