This article describes what information is needed when contacting technical support in order to troubleshoot an EMG installation.
First of all we need version and license info for EMG. You display it by running “
emgd -v“. Please send us the entire output.
Second, your configuration files. Below,
$EMGDIR refers to the EMG installation directory, usually
- All files in
$EMGDIR/dbconfig/lastok, if you are using
- All users and routing files (
$EMGDIR/routingor others in use).
Third, the log files from
$EMGDIR/log. If the problem can be recreated in a separate development installation, or emgd can be stopped anyway for a short period, please prepare the necessary log files following the steps below. For a production system, you should be able to run “
emgd -reload” after modifying the log level(s). In case you have a larger number of connectors, you may also need to limit the number of connector and pdu log files when creating the tar file.
- Stop emgd, “
emgd -stop“. Make sure there are no emgd processes still running using the “ps” command.
- Move away or remove the current log files.
- Add NOLOGSERVER and LOGLEVEL=DEBUG2 (or simply DEBUG if you are using EMG 7) at the top of your server.cfg.
- Optionally, set ROTATELOGS=100M:10 in your server.cfg. This keeps 10 generations of each log file, letting them grow to 100MB each. You can keep more files if you want, but 100MB is a good size when debugging.
- If you have many connectors (>10), instead add CONNECTOR_LOGLEVEL=INFO at the top of your server.cfg, and LOGLEVEL=DEBUG2 on the connectors that are involved in the current problem. For normal traffic, this means the connectors where the message arrives and where it is supposed to be sent out. For delivery reports, it also means the connector where the delivery report arrives, which may be different from the one where the message leaves, and the one where the delivery report is sent back, which is usually the same as where the original message arrived.
- Start emgd.
- Run the failing procedure.
- Stop emgd, “
- Make a compressed archive with the relevant files:
tar cvf logs.tar log/
gzip logs.tar“, or “
bzip2 logs.tar” if bzip2 is available on your system.
Fourth, any other files you think is relevant.
Make sure that you do not do ANY changes to the files before sending them to us. Also make sure no character mapping is applied to the files (DOS<->UNIX line termination etc). The best way is usually to put everything into different files, create a tar-archive containing these files, compress it and then send us the compressed tar-archive as an attachment. Combining everything into a single text file typically makes troubleshooting more difficult.
Send the log files and the configuration files as attachments to firstname.lastname@example.org. If you have an open ticket for this particular problem, please send the message as a reply to an email in that thread. If it’s a new problem, please make sure you send a new mail, and do not use the Reply function.
If the compressed log files are large (> 5MB), you can open up a ticket first by sending an e-mail to email@example.com and then send the files via Sprend, with the same address as recipient. Please specify the ticket id in the message.