Expiration of expired headers
Expiration of expired headers
I'm a bit confused on how UE expires headers that are no longer on the server. I have the retention set to "Natural" instead of expiring after X number of days in the hopes that UE will see what the available header range is and delete any that fall out of that range. Is there an option somewhere else that needs to be set for this to occur?
it is described in UE.txt
UE cannot know the exact time when headers are expired on servers unless you use the article number newsgroup destination type, but the type is expensive to use in the terms of RAM, so if you need to deal with large amounts of headers the solution is to set desired retention in properties->newsgroups, it is also good to set last headers to reduce the time needed during get headers for skipping headers beyond of the retention range.
during downloading headers retention is not used to expire headers so it will be the same for all newsgroup types (in principle it could be done for article number groups). for downloading headers it is only used not to download headers beyond the retention range.
last headers are only used during downloading headers, for all newsgroup destination types it will limit header range to last headers, but it only can delete older headers for the article number destination type since for other destination types the article numbers are not known.
after you set retention you can use the automatic expiration option (properties->general->apply retention automatically) or to invoke expiring headers manually ("apply retention" from the context menu in workspace).
the reason for all that article numbers are not the same for different servers, so if you have several servers, article numbers are something which requires RAM proportional to the number of servers, no problem for text or maybe pictures groups but for something much larger there is significant overhead.
moving article numbers to the disk is possible but messy, i estimate the time required for the development would be comparable with e.g. adding indexing and many operations with such a destination type would require operations like 'compacting' of all group secondary storage data, most importantly it would happen for every header task, so such a type cannot be implemented effectively.
but read help notes, i believe there is enough material there.
UE cannot know the exact time when headers are expired on servers unless you use the article number newsgroup destination type, but the type is expensive to use in the terms of RAM, so if you need to deal with large amounts of headers the solution is to set desired retention in properties->newsgroups, it is also good to set last headers to reduce the time needed during get headers for skipping headers beyond of the retention range.
during downloading headers retention is not used to expire headers so it will be the same for all newsgroup types (in principle it could be done for article number groups). for downloading headers it is only used not to download headers beyond the retention range.
last headers are only used during downloading headers, for all newsgroup destination types it will limit header range to last headers, but it only can delete older headers for the article number destination type since for other destination types the article numbers are not known.
after you set retention you can use the automatic expiration option (properties->general->apply retention automatically) or to invoke expiring headers manually ("apply retention" from the context menu in workspace).
the reason for all that article numbers are not the same for different servers, so if you have several servers, article numbers are something which requires RAM proportional to the number of servers, no problem for text or maybe pictures groups but for something much larger there is significant overhead.
moving article numbers to the disk is possible but messy, i estimate the time required for the development would be comparable with e.g. adding indexing and many operations with such a destination type would require operations like 'compacting' of all group secondary storage data, most importantly it would happen for every header task, so such a type cannot be implemented effectively.
but read help notes, i believe there is enough material there.