Guys, Gals, and Robots,

We had a brief issue with the log-in system. The fault was MediaWiki’s—apparently a PHP function was removed in their SVN for some undocumented reason, and for a reason neither Shane nor I can fathom. However, we have fixed MediaWiki’s FUBAR and the single user login should now be working, aside from the minor issue of the cookies. To work around that, please remember to check off the “Remember using this computer” before logging in!

Please shoot us any questions, comments, or concerns either via this blog or at our e-mail address.

Good hunting!

A mistap on the MediaWiki developers removed and important function that works for Centeral Auth. I re-added the function, but I will make sure that this doesn’t happen to anyone else.

As the title says, we’ve got a few updates.

Anyway, the reason why access may be lagging is because we’re testing out a new method of doing database backups, as well as some redesign of the database behind the scenes. (There are a few reasons for this, and now that we have an expert on our staff who knows MySQL like the back of his hand—Michael Casadevall—things will definitely be more optimized. However, before we can get to that plateau, there’ll probably be some growing pains.

So, if you ran into any problems for a two hour stretch from 1-3 A.M., now you know why.

More at… whenever.

I call this the great crash since it was during the Hub. In any case here is the time line of events:

1. The Hub Aired
2. The Hub Ended
3. You (That’s you reading this post connect to site) times 10000000 people.
4. Me and Joe = We try to save some sanity.
5. I do something. The database indexes and data get corrupted.
6. With you all still visiting the website, have trouble testing fixes after doing some imports. So I rebooted the servers.

Anyway. Everything should be back up in running. All sites should be working and back as new (Media wiki type of course) from June 2. So we lost 5 days worth of data. I tried to recover it, but that was no good. So that’s the deal. This effected every wiki, however, no hard data files were lost on the media wiki.

Forums are up 100%.

Right not if you do not select the “Remember my login on this computer” when logging on, the cookies disappear and you “logoff” as soon as you hit the refresh button or goto a new page. I have sent the error to the development team. This is a temporary fix for now until the normal process of “Login” and the cookies expires after the default time are back to working again!

After some fiddling around in the early hours of Friday (it’s 1 AM as I write this), Shane and I managed to get Single User Login (SUL) working.

The idea of SUL is that you can log into all Battlestar Wikis (i.e. all languages and the Media wiki) by only logging in once. To enable SUL for you, do the following:

1. Log out
2. Log back in while keeping “Remember me on this computer”. (Check this post, as this is a temp fix.)
3. Go to Special:MergeAccount. You should see a big header saying “Login unification complete”. If you don’t, you’ll have to enter your password in the form and click “Confirm login information”, then click “Unify accounts”. After that, you should see “Login unification complete” as well.
4. Go to a wiki where you’ve never even *been* before ( the Chinese wiki is a good candidate for most people).
5. You’re logged in already! You don’t even have to create an account there: it’s been automatically created behind the scenes.

Note that steps 1, 2 and 3 only have to be taken once. If you have any problems using SUL, please report them at the Chiefs’ Noticeboard and I’ll look at them. If you are having trouble, lease send us an email at webmaster [at] battlestarwiki [dot] org as you can not post as a guest.

Until confirmed by Catrope global login is somewhat working. I believe it’s have trouble moving the user accounts across. He will be looking into this. Otherwise everything else went fine this morning. To use the “global account” system you might have to totally clear your cookies from all the wikis so it works correctly.

We have decided to go ahead and update all the wikis to 1.13 before it’s scheduled release (which will be next month) in favor of being able to install an extension that will allow all of the language wikis to use one central login database.

This means there will be one “Shane” account that I could now use at en, de, zh, with only loging in once? Always tired of those creating mulipuple accounts? This is what this solves. Not only are we going to extend it just for the Battlestar Wiki, but it will extend to two other related sites. Battlestar Pegasus and Wiki Frakr!. Those will not be “active” until 1.13 is in official release status and depending on how well it works out, we will expand it’s usage to also the Eureka Wiki.

So that’s the deal. If you see any site problems in the next few hours, this is why.

Today we had a very successfully maintenance day on the wiki’s and the servers all at the same time. (Go figure…)

In any case, if you didn’t notice already, the servers never went down. That’s because the servers where not to go down, but the backend for us to access their support system was down. Go figure (as I said before.) So this is done.

After many many months of the Task Perl Bot not working, it now is working and will start automatically updating the open task list on the English wiki, thanks to our newest bot installed, Media Tag Bot (By Carnildo). I used the new perale backend code to re-code the Task Bot, and it worked fine. A few errors, but working those out with Carnildo.

The Media Tag Bot examines all recently updated media files and checks to see if they are tagged with the correct copyright tags. If not, it will notify the user and tag the media file according to the rules we have set forth in it’s logic based upon the set polices on the Media Wiki. This will help us make sure we are covered legally and have no issues in the future.

If you have any questions on what went on in the past 48 hours, just give us a shout out on the [[BW:CN|cheif's board]] and we will get back to you!
So say bot us all!

Our data center is making upgrades to the infrastructure on May 10th, 2008. When the servers are down they estimate that it will be as long as eight hours our more. We hope it’s a lot less time. However, we have two servers on two different areas of their data center.

Since there is nothing we can do about this, please check the Case Orange blog, itself on another server, for updates. We are getting this message out now so you don’t think we have disappeared on you!

If you have any questions, post them here.

I am unsure how long the MySQL Database was down in the middle of the night, but it’s back up and running. I am looking into this since the MySQL server was running when I checked the services page.

All,

In the event that you may have noticed, our Athena database server was briefly down. That’s because I had to have it rebooted, though it had nothing to do with performance reasons: I accidentally shut down SSH and couldn’t get back in, so I had the Planet people reboot the server.

Hence the slight hiccup… which you probably didn’t notice.

I thought I would mention it anyway.

Back to burning the midnight oil…