A few days ago I installed UE 3.3 trial 32 bits in my Windows XP (32 bits) SP3 with 3 GB RAM. So far all features I tried are great (use of boolean wildmats, filters, search index, collections, auto download/par/unrar) but I got a little problem that is bugging me alot. I can subscribe and get headers from 2 groups containing 800 000 and 700 000 headers in about 140 MB of RAM but when I try to download headers from another group that contains 9 000 000 I got a database too large error (Task Manager shows UE using above 1 GB of RAM when theres about 5-6M headers get). Reading the forum I tried unchecking edit menu->properties->general->database too large to fit in memory, but now the program just quit (not even shown in task manager). The groups are created default (compact binary), retention is 1200d as my provider NewsHosting got almost that, headers per server is 30 000 000 (defrault).
As I read in the forum Its supposed not to have any problem to get 9 000 000 headers in my system but I just dont know what else to try, please give me some advise.
UE database too large error when getting 9 000 000 headers
-
JackieChun
- Posts: 5
- Joined: Tue Nov 15, 2011 5:16 pm
Re: UE database too large error when getting 9 000 000 heade
32 bit system has RAM limitations, 64 bit process can utilize all the RAM available.
How many headers strongly depends what is in the newsgroup.
If the newsgroup contains large amount of spam (if you select all headers with ctrl+a you'll a number of entries of the same order as the number of headers), it takes a lot of memory since there is much less compression. You can see all that spam when scrolling, e.g. if you purge (not just delete) those spam batches will be more RAM available.
Or e.g. picture group with the same amount of headers will take much more RAM comparing to binary group containing large rar files.
Basically the amount of RAM is proportional to the number of entries in the newsgroup, you see the number in the title bar when you select all entries in the newsgroup.
Some groups were and are being heavily spammed, providers don't do anything about it.
How many headers strongly depends what is in the newsgroup.
If the newsgroup contains large amount of spam (if you select all headers with ctrl+a you'll a number of entries of the same order as the number of headers), it takes a lot of memory since there is much less compression. You can see all that spam when scrolling, e.g. if you purge (not just delete) those spam batches will be more RAM available.
Or e.g. picture group with the same amount of headers will take much more RAM comparing to binary group containing large rar files.
Basically the amount of RAM is proportional to the number of entries in the newsgroup, you see the number in the title bar when you select all entries in the newsgroup.
Some groups were and are being heavily spammed, providers don't do anything about it.
-
JackieChun
- Posts: 5
- Joined: Tue Nov 15, 2011 5:16 pm
Re: UE database too large error when getting 9 000 000 heade
Thanks Alex, that was the problem. Now I unchecked auto/custom collections from the toolbar and also unchecked show threads (toolbar). I got: 373 569 and 318 193 entries (items) in each group. They have about half the entries than the number of headers in each group: 800 000 and 700 000 headers. So I think that 5M entries/items take about 1 GB of RAM, is that right?
I can purge the spam manually, but how to purge automaticaly the auto download&saved attachments?
I can purge the spam manually, but how to purge automaticaly the auto download&saved attachments?
Re: UE database too large error when getting 9 000 000 heade
no need to purge saved attachments, they are just marked read/deleted depending on settings, so UE will remember them when rescanning headers.
you can purge deleted headers, but no reason to do that only for saved attachments.
you can purge deleted headers, but no reason to do that only for saved attachments.