On Jan 31, 2019, at 4:32 AM, Evgeniy Kuznetsov < evkuz@xxxxxxx> wrote:
Hi all,
Our team using virtual machines as execute nodes and all of them doesn't have public ip address and FQDN, so the submitter cann't receive a notification email about job completion.
Error message looks like this :
"host mail.jinr.ru[159.93.14.110] said: 450 4.1.8 <condor@xxxxxxxxxxxxxxxxxx>: Sender address rejected: Domain not found (in reply to RCPT TO command)"
Hostname "wn_221_141.jinr.ru" is not a FQDN, it is specified in /etc/hosts.
Is there any way to tell condor use some login/pssword pair for execute node to connect to mail server as a client and send email from dedicated e-mail address ?
First thought I had was you can set the SENDMAIL config knob in condor_config to tell HTCondor to run a custom script whenever sending email instead of simply invoking /use/sbin/sendmail. Your script could do whatever is necessary to send out an email in your
environment. The message and destination will be passed to your script via stdin, just like invoking sendmail (see âman sendmailâ). Below is copied from the HTCondor manual.
Cheers,
Todd
SENDMAILThe full path to the sendmail executable. If defined, which it is by default on non-Windows platforms, sendmail is used instead of the mail program defined by MAIL.
MAILThe full path to a mail sending program that uses -s to specify a subject for the message. On all platforms, the default shipped with HTCondor should work. Only if you installed things in a
non-standard location on your system would you need to change this setting. The default value is $(BIN)/condor_mail.exe on Windows and /usr/bin/mail on all other platforms. The condor_schedd will not function unless MAIL is
defined. For security reasons, non-Windows platforms should not use this setting and should use SENDMAIL instead.
|
![[Computer Systems Lab]](http://www.cs.wisc.edu/pics/csl_logo.gif){kind=logo}