amuck-landowner

Forum downtime / maintenance RFO

MannDude

Just a dude
vpsBoard Founder
Moderator
Yeah, lots of posts no appearing in public.

How you are able to get it figured out and documented for long term sanity.
The data still exists, so it's not lost. Just not viewable. I'm waiting for IPB to get back to me, as I'd like to manually restore any lost discussion that's not viewable currently.

None of the posts have been lost so don't worry. It's just IPB and its archive feature that appears to have 'removed' them from thread visibility.

They call that a "feature"?  That isn't a feature.   

Why would anyone running a site want to see their posts disappear from their site?
Well, the feature is to archive things based on certain criteria. The settings were incorrect (my fault) so it archived damn near everything. While UNarchiving posts, it won't complete. Then the CVPS thread (and that thread only, I believe) has lost half of it's viewable content. The content still exists, just not viewable now...

Further more the daily post stats are screwed up in the backend and don't represent the true daily post counts anymore.

I could run some DB tools from the backend to rebuild/recount/redo stuff but waiting to hear back from IPB before I do anything.
 

mojeda

New Member
@manndude, you can export mysql database remotely via ssh as long as the server has mysql-client installed :)

mysql -h mysql.host.tld -u USERNAME -pPASSWORD DATABASE >> database.sql

Something like that at least, I don't see why this wouldn't work with BuyVM's Shared SQL.
 
Last edited by a moderator:

MannDude

Just a dude
vpsBoard Founder
Moderator
@manndude, you can export mysql database remotely via ssh as long as the server has mysql-client installed :)

mysql -h mysql.host.tld -u USERNAME -pPASSWORD DATABASE >> database.sql

Something like that at least, I don't see why this wouldn't work with BuyVM's Shared SQL.
We did this yesterday morning. The issue isn't importing the DB now, the issue is pulling specific data from it and restoring some posts individually.
 

JDiggity

New Member
Forum was down for several hours while Martin and I got everything sorted. In short, we had to restore a DB from about 5AM EST, I put the site in matineance mode around 6AM(ish) so it was a slow period, not many (if any) posts were lost.

Chat log from IRC because I am tired:



11:40 <@Manndude> Well here is the series of events


11:40 <@Manndude> Saturday night, MannDude had some beers


11:40 <@Manndude> NOT A LOT


11:40 <+TheHackBoxPie> oh crap


11:40 <+TheHackBoxPie> he franned it


11:41 <@Manndude> The post stats in the backend were broke. They were displaying the wrong numbers for the days. It was saying that prior to the 18th the forum was getting liek 40 posts a day. The 18th we had 777 posts exactly, I remember because it was 777


11:41 <@Manndude> Soooo


11:41 <@Manndude> I do a DB dump, and THEN run the optimize tables tool


11:41 <@Manndude> Things were faster, but I noticed some threads were archived that should not have been


11:42 <@Manndude> So I tried to fix that


11:42 <@Manndude> By turning off archiving and setting it to UNarchive anything


less than 365 days old (everything)


11:42 <@Manndude> then I noticed the CVPS thread went from 700+ posts to 350 something


11:42 <@Manndude> So locked her down


11:42 <@Manndude> Said 'nuh uh, not today'


11:42 <@Manndude> And started drinkign coffee


11:43 <@Manndude> But the DB dump was too big for PMA which is my only point of access to the remote server and doing it via SSH took a long time


11:43 <@Manndude> plus there was an hour or two trying to figure out how to proceed without doing a DB restore


11:44 <@Manndude> That's the RFO


 


As of now, 11:50 GMT-5, the board is currently unarchiving threads and will be doing so for some time.
attachicon.gif
Screenshot - 06232013 - 11:59:14 AM.png

That is the screenshot of the post stats now. That is completely wrong and much lower than what it should be. The 18th we had exactly 777 posts.

I believe this has to do with all the thread archiving. I enabled archiving but didn't specify the proper settings, thus damn near everything archived. I goofed, it was my fault. I'm paying the price for it now at noon with no sleep. Everything should be unarchiving now.

The way this works is there is a scheduler that runs on it's own unknown (to me) interval that does this in bulk, 250 at a time. It looks like it took quite some time for everything to archive, so I suspect it'll take quite some time for everything to UNarchive.

That's the RFO for now.
If you set the upload file size larger and the php timeout to a longer time then this should not be a problem.
 

MannDude

Just a dude
vpsBoard Founder
Moderator
I believe it was this part:
MySQL is offsite/remote on a shared SQL server. I can not adjust those settings myself. Had to import via SSH which for some reason was unreasonably slow. Took several hours for a 90MB DB, and the import failed the first time as the connection dropped mid-way.
 
Top
amuck-landowner