piler import
Hello,
I started pilerimport from folder with more that 1 million of eml and after import most of them the import command have been aborted. Does pilerimport look for a eml by date or by name? I have only last 20 names of imported eml and it looks like name used in random way. I do not want to waste another 24 hours to import duplicated eml.
Thanks.
Comments (12)
-
reporter -
repo owner So in this case we should use scandir() rather then the usual readdir(), right? (Since readdir() provides nothing to tell it to sort by something).
-
repo owner -
assigned issue to
-
assigned issue to
-
reporter Another issue we found. Affected build 857, 866.
After pilerimport from mbox or maildir two fields are empty in metadata and sph_index: from, fromdomain. Piler_id saved as message_id in metadata. All imported messages had Message-id.
-
repo owner Just tested the with build 866, and got the following results, see the attached files.
-
repo owner - attached p1.txt
-
repo owner - attached p2.txt
-
repo owner Can you show me a few problematic messages? I'd try to import them too, and see what happens.
-
repo owner Ehh, I've just removed the attachments, don't let those email addresses fall in to the hands of spammers. I may send them to you in email.
-
repo owner Thanks for the samples. Exchange gives me a headache. What version do you use? I tested piler with exchange 2007 sp2, and it adds a journaling header, and puts the original message as a message/rfc822 MIME part. However your exchange behaves a bit differently.
-
repo owner Not sure if the problem still persists. If so, then try it with the latest master branch as well, and reopen the issue, if it didn't fix it.
-
repo owner - changed status to resolved
- Log in to comment
As i understand readdir will return results similar to random. http://utcc.utoronto.ca/~cks/space/blog/unix/ReaddirOrder Can you sort input files by date or name. This can save a lot of time when import aborted and need start again.