What if we got rid of all of the old forum archives? Only keep old posts for a year (2000/01)? I mean, who actually goes and reads the really old posts in there? Wouldn’t that save a lot of space on whatever it is that runs the board? I don’t know maybe it is a bad idea and I should just be quiet.
It’s really an excellent suggestion, but they have many times refused such suggestions.
One good one was to delete everything in the pit over 2 weeks old. That would be no loss.
Things in MPSIMS, likewise, are not worth repeat looks after 2 weeks.
I’ve often wondered this too-- of course GQ should be kept around in archives, as it is sort of the original point of the board, but is there really a need to keep posts about tonight’s episode of Friends and how I just grazed my knuckles on a stove burner for future consideration?
Buuut … My finest posts are in the Pit!
Why have they refused the idea?
I agree ther everything in the pit, MPSIMS, even IMHO over two weeks old should should be removed.
For what purpose? What’s it to you if the old posts remain extant?
The size of the board affects everything about it.
Maintenance is slower, disk usage is greater, the newest posts are farther out on the spindle for longer seek times, etc.
It reaaaly slows searches, even if you specify “last 30 days”.
This is because the index is not sorted by date, but alphabetically.
The 30-day part only reduces the number of answers returned, not the work to locate them.
OK, so we ARE back to talking about speed issues. (That wasn’t apparent from the OP or the posts that followed).
I would be happy to see everything older than 30 days spooled to a different (read-only) database. That would speed up the hot & active database but would still preserve the old posts for searching and we could still link to them in new posts.
I would be very unhappy if the old posts were simply deleted, though.
Just curious, but where do you get this information?
Seems like there might be 500 “test” posts which seem absolutely useless to me. Isn’t that why we have a preview button? Toss all them too. They’re really just MPSIMS stuff in ATMB.
My question didn’t really have anything to do with speed, it is just something that came up.
Well, if it isn’t for reasons pertaining to speed, it can’t reasonably constitute so much as a hair up anyone’s nether regions if the archives continue to exist.
And even if speed IS an issue, there are other ways of separating the currently active threads from the archives, such as the one I’ve already suggested here.
Put this digital pack-rat down for a resounding and emphatic “NO” to the proposal that we kill old threads.
Every now and then I go back and delete old test threads. I haven’t been doing it lately, because I haven’t known just what we’re gonna do with the message board. I’d say that I’ve deleted a couple of HUNDRED test threads in ATMB, and probably at least that many threads reporting problems. Back in the UBBS days, the board would go down just about every weekday in the middle of the afternoon, and people would report it in this forum.
This seems highly unlikely. It is a relatively simple matter to index the “LastPost” field in the database engine, which allows the use of an index slice instead of a full scan.
Your supposition that “the newest posts are farther out on the spindle” lacks merit given that the Boards run on a Linux machine using MySQL as a database engine. Linux does not use the dumb first-free-block allocation algorithm that Windows uses, and the MySQL database engine’s use of disk space is complex, to say the least. (Does vBulletin store the posts in the database engine, or just pointers to the file system? I’ve never used the software.)
Has anyone tried to graft glimpse onto vBulletin? That would really speed up searches. I’d be willing to help develop a custom system for the Boards provided the end result is open source.
All dynamic content, such as posts, are stored in the database. The Posts.myd file becomes quite large.
I say we dump stuff over a month old, or transfer it to a different site.
Is there really any good reason to keep 682 pages of General Questions around?
Or 609 pages of MPSIMS?
123 pages of ranting from The Pit?
54 pages of About This Message Board?
Etc., etc., etc.
All told, it adds up to a staggering 1,862 pages of spanners in the works, gumming up the speed of the board.
We don’t need it.
Weed it!
IIRC the amount of space the old posts (all of them) take is trivial in terms of disk space. I think it was quoted as being a few gigs at best and with fast 40 gig drives being sold for $ 150 I don’t think this should be an issue re storage. The issue (I think) is how fast the DB engine and related systems can respond parsing through this quagmire.
I think the old posts are quite valuable and should have a special and separate place on the board with their own search engine front end that will not affect the speed of the main SDMB DB engine.
Doing this will cost time and money. The Reader is probably not inclined to indulge this board’s needs and wants much further without seeing a return of some kind.
In my professional opinion, they shouldn’t make that much of a difference. I am quite familiar with the database engine they’re using and it should not degrade significantly under that sort of usage, assuming that the proper indices have been set up on the tables in question and that the software avoids making unnecessarily expensive queries.
An upgrade to MySQL 3.23.23 or later and using the FULLTEXT index type might improve search performance dramatically, but this is probably nontrivial.
Sorry, the suggestion of “getting rid” of old threads in GQ (and CCC/CSR) is absurd. These are an immense resource of information. It would amount to an act of vandalism against a unique source of knowledge. The very suggestion turns my stomach.
I’d agree that some of GD, and most of MPSIMS and the Pit is expendable. However, there’s a lot that is valuable there as well - e.g. some of the race threads in GD provide a fuller discussion than anything in GQ. There are some very funny threads in MPSIMS and even the Pit that it would be a shame to lose. But sorting the wheat from the chaff would be very time consuming - a better solution would be to archive everything: Possibly older than two years for GQ and Comments, and three months for everything else.
The total number of deleted threads is currently about 5,577.
Based on the main menu page saying “93116 threads in total” and the most recent thread #98693.