Page MenuHomePhabricator

BadMethodCallException with "Call to a member function getId" in SpecialCentralLogin.php on Beta Labs
Closed, ResolvedPublic

Description

Upon going to Special:UserLogin on English Wikipedia on Beta Labs (http://en.wikipedia.beta.wmflabs.org/wiki/Special:UserLogin), I got redirected to Special:CentralLogin/complete (with token). However, it then threw an exception:


Unexpected non-MediaWiki exception encountered, of type "BadMethodCallException"
[31c07ff4] /wiki/Special:CentralLogin/complete?token=201d1e72216aa80eef3f599b689e7bdc Exception from line 166 of /srv/common-local/php-master/extensions/CentralAuth/specials/SpecialCentralLogin.php: Call to a member function getId() on a non-object (boolean)
Backtrace:
#0 /srv/common-local/php-master/extensions/CentralAuth/specials/SpecialCentralLogin.php(33): SpecialCentralLogin->doLoginComplete(string)
#1 /srv/common-local/php-master/includes/specialpage/SpecialPage.php(357): SpecialCentralLogin->execute(string)
#2 /srv/common-local/php-master/includes/specialpage/SpecialPageFactory.php(516): SpecialPage->run(string)
#3 /srv/common-local/php-master/includes/MediaWiki.php(300): SpecialPageFactory::executePath(Title, RequestContext)
#4 /srv/common-local/php-master/includes/MediaWiki.php(609): MediaWiki->performRequest()
#5 /srv/common-local/php-master/includes/MediaWiki.php(460): MediaWiki->main()
#6 /srv/common-local/php-master/index.php(46): MediaWiki->run()
#7 /srv/common-local/w/index.php(3): include(string)

#8 {main}


Version: master
Severity: major
See Also:
https://bugzilla.wikimedia.org/show_bug.cgi?id=69834

Details

Reference
bz69565

Event Timeline

bzimport raised the priority of this task from to Unbreak Now!.Nov 22 2014, 3:31 AM
bzimport set Reference to bz69565.

Confirming.
No idea if this is specific to BetaLabs or a general issue in CentralAuth, hence tentatively setting highest prio and CC'ing Greg and Chris.

We ran into this on testwiki (I2488e1fd3109ce66c2cb42d9afc78097f2b7313d) which Reedy backported and deployed so now it's throwing a MWException instead of a fatal error. I'm not sure whether it is supposed to be redirecting the user in the first place though...

Are we running low on session/cache storage (I'm not sure if it's redis or memcached backed in beta)?

I can login fine there-- I'm assuming this is intermittent?

(In reply to Chris Steipp from comment #4)

I can login fine there-- I'm assuming this is intermittent?

I tried now ~15 times (bypassing Firefox' browser cache) and it showed the exception every single time.

Hmm. Browser specific bug?

I can login with Chrome without any problem. Firefox I don't get this exception, instead I (repeatably) get:

Function: EchoTargetPageMapper::fetchByUserPageId
Error: 1146 Table 'enwiki.echo_target_page' doesn't exist (10.68.17.94)

(In reply to Chris Steipp from comment #6)

Hmm. Browser specific bug?

I can login with Chrome without any problem. Firefox I don't get this
exception, instead I (repeatably) get:

Function: EchoTargetPageMapper::fetchByUserPageId
Error: 1146 Table 'enwiki.echo_target_page' doesn't exist (10.68.17.94)

Scratch that. I get this ^^^^ error for every single article page when I'm logged in, with both browsers now. After Echo fixes their issue, I'll keep working on this bug.

(In reply to Chris Steipp from comment #7)

(In reply to Chris Steipp from comment #6)

Function: EchoTargetPageMapper::fetchByUserPageId
Error: 1146 Table 'enwiki.echo_target_page' doesn't exist (10.68.17.94)

I get this ^^^^ error for every single article page when I'm logged in, with
both browsers now. After Echo fixes their issue, I'll keep working on this bug

Has the Echo team been notified (CC"ing some more folks now)?
Is there a dedicated bug report about that issue?

(In reply to Andre Klapper from comment #8)

Has the Echo team been notified (CC"ing some more folks now)?
Is there a dedicated bug report about that issue?

Ah, might be bug 69590

(In reply to Andre Klapper from comment #9)

(In reply to Andre Klapper from comment #8)

Has the Echo team been notified (CC"ing some more folks now)?
Is there a dedicated bug report about that issue?

Ah, might be bug 69590

That's "fixed" (waiting on proper solution, but symptoms should be gone).

Chris S: retest?

I think this is fixed by I2488e1fd3109ce66c2cb42d9afc78097f2b7313d (the change Kunal mentioned).

I think the underlying issue is bug 69834.