Having migrated my scholar site from beta-13 to RC1 I was happily working away, cleaning up my list of important biblio entries. I had switched from sorted-by-year to sorted-by-author. I selected the pull-down option to delete an entry that Google scholar has put in in error. As soon as I invoked the delete operation, I received an "internal server error 500 code." I don't understand the back end well enough to know which logs to check. Re-running the Aegir validation suite for the site *fails* however, with the following two errors:

(1) Running: /data/disk/wizard/tools/drush/drush.php @kevinjsullivan.org provision-verify --backend 2>&1
The external command could not be executed due to an application error.

That error occurs just before the line "Bootstrap to level 0" in the Aegir validation output.

(2) Drush command terminated abnormally due to an unrecoverable error. Error: require_once(): Failed opening required './profiles/openscholar/modules/contrib/purl/includes/purl_processor.inc' (include_path='.:/opt/local/lib/php:/usr/lib/php5') in /data/disk/wizard/static/platforms/openscholar-2.0-RC1/sites/all/modules/contrib/ctools/includes/plugins.inc, line 759

Digging a little shows that I do have a ./profiles/openscholar directory, but (!!!) there is no modules subdirectory here, within which the code is looking for ....contrib/purl/includes/pur_processor.in.

I've marked this error as critical, as it clearly completely crashes a site. I can no longer access any page. Attempting to do so just yields internal server error 500.

Note: I followed the standard Aegir procedure to migrate my 2.0-beta-13 site to 2.0-RC1. I cloned my beta site, validated the clone, migrated the clone to 2.0, validated the result, worked with it for a while to confirm that it seemed ok, then migrated my live site to RC1, worked with it for a while, and then (to my chagrin) deleted the old clone. Then this problem occurred, and now I'm stuck unable to access my old content. It would be wonderful if someone could help me to fix this problem and recover the site that I've been working on now many hours!

Regards,
Kevin

Comments

kevinsullivan’s picture

kevinsullivan commented
Priority: Critical » Major

Well, I'm downgrading this bug to "major" in that it's impact is major but I'm guessing that's maybe just a consequence of still being in beta/RC. I've tried a bunch of thing to fix this problem, including (1) running drush commands from the command line -- I cannot execute a single meaningful drush command against this site, getting the same error as indicated above; drush fails immediately; it clearly cannot "boot up" to do anything with this site, due to some kind of configuration issue; (2) clearing cache_ variables through MySQL (MariaDB); (3) doing a rebuild_registry (which also fails, producing a headers already sent error); (4) clearing what I expected to be a registry table in the DB -- but there isn't one. I've tried to track down the source of the non-existent path by searching the source code for the entire project. It must be in my database somewhere, because I can't find even relevant parts of the path in any of the code. Also, creating a new site on the same platform works fine. Diff'ing the two sites reveals many differences (even though they're ostensibly running on the same platform), but nothing jumps out as being the immediate cause of the problems I'm seeing. There's obviously something fundamental that is corrupted in the broken site, probably simple, but I just haven't been able to find it, not being a Drupal or PHP wiz. And I've run out of time for debugging. I've done an SQL dump and will recover writings that I'd had in my site from that text dump.

kevinsullivan’s picture

kevinsullivan commented

I've re-installed 2.0-RC1 into Aegir. The platform validates. I created a new scholar site. It validates. I added a scholar web site. I configured its layout. All was fine. I add a class. I fill in the form. I hit save. Error 500, internal server error. I'm losing my patience :-( The Aegir environment is a clean build by boa / barracuda on an EC2 VM running an Amazon Ubuntu 12 AMI and with a satellite aegir instance installed by octopus. It really seems like this should work out of the box. Unfortunately, the system that barracuda builds, based on nginx, has really minimal logging. It's hard for me to see what's going on.

ferdi’s picture

ferdi commented

This is a bit tricky without some more info / error logs because we dont use Aegir. You probably can get some help at #openscholar (IRC) ?

thanks!

niccolox’s picture

niccolox commented

hi kevin

I am using the latest BOA 204 dev and which has OpenScholar RC2

OpenScholar install OK but the auto-site creation fails and much of the theming and OpenScholar spaces/context/features presets is lost