I've trawled through some old funnelback error emails, and found that one collection on one server did produce details from logs in the email, while all others don't. Looks like it was during the index phase. I wonder if it's something specific to that setup? Here's what it looked like:
HOST: xxx-funnelback01.squiz.net
COLLECTION: xxxx
MESSAGE: Error: Error running "/opt/funnelback/linbin/ActivePerl/bin/perl-static /opt/funnelback/bin/index.pl /opt/funnelback/conf/xxxx/collection.cfg".
UPDATE LOG
...
CRAWL LOG
...
MESSAGE end
I've tried using a bash script to return an error code on the following workflow commands, but none of them cause logs to appear in the failure email:
- post_gather_command
- pre_index_command
- pre_swap_command
Would it make any difference if I were running a groovy script? I can't see how it would. I suppose this just doesn't work for workflow commands.
Anybody able to shed some light?