Today, on our mail relay several sendmail processes continued logging
STARTTLS: read error=generic SSL error (0)
in an infinite loop until until our log filesystem was full. This was the
first time after running for about one month under heavy load without
problems. All processes have been started by connections from one single
SMTP-Client system outside of our organization.
We are using sendmail Version 8.12.5 on a RedHat 7.3 Linux System with rpm
I have looked into the source and detected a possible reason for the
infinite loop in tls_read.c . In some cases the count variable "again"
(I guess it should protect from such loops) will be reset and the message
above will be logged.
In the current release 8.12.7 there have been changes in this part of the
Does _FFR_DEAL_WITH_ERROR_SSL cure our problem?
Is this Define enabled by default?
With best regards - thanks for any kind of advice
Walter Mller - Zentraler Informatikdienst - Universitaet Innsbruck