[e-mail]

[Marcel Bischoff]

Since setting up Piler 1.2.0 from scratch, I get quite a lot of those "messages" in the list:

Selecting them does not display anything. Downloading the EML file results in the following inside of the file:

if allof (spamtest, not address :all :is :iscontact "From" "") {
  fileinto "Junk E-mail";
  stop;
}

[Janos SUTO]

This looks like a sieve rule. You should feed pilerimport only with emails, not other files.

[Marcel Bischoff]

This is probably due to Kerio Connect storing all kinds of metadata inside the message folders. I really don't like this software. It's not doing things the way they are supposed to be.

How would I clean up those emails programmatically with a one time command?

[eXtremeSHOK]

@herrbischoff

Im pretty sure I can write a custom script to correctly export the kerio connect emails and then mass import them to piler.

[Janos SUTO]

The parser skips files shorter than 100 bytes. Unfortunately this file is 101 bytes just passing this test. However, it's probably true that any emails under 200 bytes are not real emails. To check this hypothesis run the following query:

select id, piler_id, from, subject, size from metadata where size<200;

If you conclude that they are indeed not emails, then set the retained column to a timestamp in the past (eg. yesterday), and finally run pilerpurge. Note that it will purge all emails having a retained value in the past! If you set the retention to 5-10 years, then it won't be a problem. However if you have set the retention to 100 years or so, then don't do it, because a 'bug' makes such values to overflow in the mysql column. (Not sure if it's a bug anyway)

[Janos SUTO]

No news is good news.

[Marcel Bischoff]

Sorry for the late response. I have tested your hypothesis successfully. It is indeed true that no email under 200 bytes is a real email.

[Marcel Bischoff]

@extremeshok: Thanks, but that will not be necessary. I already had an export that needed to be imported.

[Janos SUTO]

OK, then it's safe to set the min_message_size variable to 200, and the next time you may skip the sieve stuff.