Using way too much Hard drive space

jimmydo2 · Post by **jimmydo2** » Wed Jul 11, 2007 12:05 am

How do I reduce the database usage?

I Have version 1.9.5

My Data folder is taking up over 16 Gigabytes, yes that is Gigabytes
16,240,337,255 Bytes

SPlit pretty evenly between the bodies folder, and the newsgroups folder (8.7 gig for bodies, and 7 gig for Newsgroups

with 400 Meg in the temp folder

THe Newsgroups folder has one folder with over 1 gig in it all by itself

Each folder in the bodies folder contains about 500 Meg.

I have run Compact and free

and Purge deleted headers

I have no files queued up in the the save Queue

Please help

Josef K · Post by **Josef K** » Wed Jul 11, 2007 2:33 am

Having that much space occupied by bodies indicates lots of articles downloaded but not saved. In Task Manager, switch to the Downloads tab and see how many (if any) articles are there. If there are articles present, it means you have downloaded but not saved them. If there aren't any there then somehow they have become disconnected from the database, maybe through a crash. The bodies directory isn't part of the database - it's a storage space for articles downloaded (separate segments) until the entire article can be retrieved before combining.

A large Newsgroups directory can mean you either haven't optimised your retention settings relative to your server or you are using a type other than compact binary. Open map.txt in your '%UE%\Database\Newsgroups' directory and check whether or not your binary and large groups are compact binary.

Have you experienced a crash recently?

jimmydo2 · Post by **jimmydo2** » Thu Jul 12, 2007 3:00 am

Unfortunately I have had a few Crashes

The downloads tab only has about 10 items in it

Map.txt lists all my Newsgroups as Regular, Compact binary.

How do I optimize my retention settings?
my Server has about 45 days retention.

If the files in Bodies have become orphaned due to Crashing, should I just manualy delete those, and the old files in the Newsgroups folder?

Josef K · Post by **Josef K** » Thu Jul 12, 2007 3:24 am

jimmydo2 wrote:Unfortunately I have had a few Crashes
The downloads tab only has about 10 items in it

Check your %UE%\Database\log.txt and either post the log here or email it to Alex. Any crashes resulting in data loss would be best looked at by him.

jimmydo2 wrote:How do I optimize my retention settings?
my Server has about 45 days retention.

In Properties->Newsgroups->Dates/Range, set the date to whatever is closest to your server, maybe a little bit over in case your NSP occasionally holds slightly more than advertised in certain groups. For example, in your case you can set '45d' or even '50d' when using compact binary since the difference is negligible, database storage-wise.

jimmydo2 wrote:If the files in Bodies have become orphaned due to Crashing, should I just manualy delete those, and the old files in the Newsgroups folder?

For now I'd say leave them and wait until Alex advises. At least save any articles that remain in your Downloads tab. Anything else left in Bodies would be strays if they aren't shown in Downloads and, unless Alex has written an articles recovery util, they can be classed as lost. In future, always make sure Properties->General->Miscellaneous->Settings backup is checked. Then you'll have several incremental versions of your settings to revert to in case of crashes.

alex · Post by **alex** » Thu Jul 12, 2007 6:33 am

to see all article bodies invoke edit menu->bodies.

then delete all you don't need, if you delete everything from there the bodies directory size will drop to almost zero.

as to the newsgroups directory you need to set some retention in edit menu->properties->newsgroups in the topmost default entry, if you set it to be the same as your server retention or higher it will be perfectly ok.

then after you open a newsgroup its size will automatically drop (or invoke workspace Newsgroups entry context menu entry->advanced->load and then file menu->save all, ensure properties->general->apply retention automatically on load is checked (it is default).

you can delete both bodies and newsgroups directories directly but when you restart the program afterwards to synchronize header ranges you need to invoke in the workspace, Newsgroups tab, Newsgroups (topmost entry) context menu->advanced->reset header ranges.

terminating UE in abnormal way cannot create any junk data, so it is something related to your settings and accumulated unsaved files over time. with crashes try to check log.txt if you see something there since I don't see anything from you here for year and half from the time when I was refining collections, so most likely you are talking about the system crashes or like (in which case you need to look for third party applications with drivers which may cause them).

jimmydo2 · Post by **jimmydo2** » Thu Jul 12, 2007 6:56 pm

edit menu->bodies, shows signifcantly fewer bodies than are stored in the bodies folder ... So I went ahead and deleted everything in the bodies sub folders that were more than 2 days old, since the way I use UE, I should not have any body parts that old. - Now the Bodies folder is just at 59 Meg.

My Retention for all Newsgroups has been set to 60 days since I set up UE

Apply Retention was not checked in my General Properties (My Bad)

After following Alex's Steps, my Newsgroups folder is now down to 3 Gig, much better thank you.

I thought the Crashes were system Crashes, No App Crashes, but I do have a Crash Log (Or will that get created for all abnormal Terminations, including system Lockups?

THank you's To Alex and Josef for all your Help

Here is my Last_Crash.log

06-Jul-07 00:37:59
UE1950R internal error (5CA668/C0000005)

alex · Post by **alex** » Fri Jul 13, 2007 3:05 am

005CA668 - the address is in the third party interface library, command bar, it is like comprising menus, workspace and task manager panes and like, statistically i didn't see such crashes at all (zero known occurences in the past), do all your crashes with this version have this address and crashes with previous version have similar address starting with 005xxxxx?

better email me your log.txt so i could check it.

jimmydo2 · Post by **jimmydo2** » Fri Jul 13, 2007 7:09 am

Well I searched my Log file, and it appears that the July 6th crash, is the only 005 error. Here is all of the Internal Errors for the last 12 months

27-Jul-06 00:44:42
UE1700R internal error (7C93426D/C0000005)

21-Apr-07 01:42:44
UE1740R internal error (4018D3/C0000005)

06-Jul-07 00:37:59
UE1950R internal error (5CA668/C0000005)

alex · Post by **alex** » Fri Jul 13, 2007 12:40 pm

i don't think it is a big deal then. if last crash is one time crash probably you hit something in the interface library, but it is not related to the original question.

in global filter, bodies you should see all the bodies, just check incomplete partials etc are shown as well.