Can't build neighbour index and systems index

[Flevity]

structures and ships build succesfuly
Building neighbour index process is stucking on wait..
Bulding systems index return an error 500
env:
clean pathfinder install
ubuntu 18.04
nginx 1.14.0
php-fpm 7.2.7
nginx_error.log

[Slivo-fr]

Same here, tried about 50 time to build system index over 48h without any success

[Derjyn]

I third this. Glad I'm not the only one, was starting to drive me mad.

Ubuntu 16.04
Nginx 1.10.3
PHP 7.1.20

[exodus4d]

I know :( there is some kind of "race condition" between the Ajax build on /setup page and a Cronjob that runs every 60s und tries to update the static eve_universe DB.

1. Please comment out L60 https://github.com/exodus4d/pathfinder/blob/develop/app/cron.ini#L60
2. Wait 2min (makes sure there is no cronjob running for update)
3. Clear cache dir or Redis DB
4. Go back to /setup page and retry building the systems index
5. Please let the cronjob (L60) deactivated until I fixed it in the next release. This is not a problem because CCP does not update static data that often

Hint: Having the system index build up is not a must have its just a performance boost. Pathinder automatically build the index for a system if it is missing in the index und requested for the first time.

[exodus4d]

If you get other problems in 1.4.0 please read this issue and check out my fix (develop branch):
#658 (comment)

[Flevity]

I commented L60 and still cant build system indexes =/ error 500 and same error on nginx log.
I've tried to login into my pathfinder and after creating corp map there is an error 500 too =/ (mapmodel.php already replaced)

building indexs error:

2018/08/04 17:12:52 [error] 1167#1167: *9 FastCGI sent in stderr: "PHP message: HTTP 404 (GET /ws/map/update)
PHP message: [index.php:23] Base->run()" while reading response header from upstream, client: 2a02:2698:4c25:5392:9968:afa8:6097:7ef7, server: achtung-wh.tk, request: "GET /ws/map/update HTTP/1.1", upstream: "fastcgi://unix:/var/run/php/php7.2-fpm.sock:", host: "achtung-wh.tk"
2018/08/04 17:13:08 [error] 1167#1167: *1 FastCGI sent in stderr: "PHP message: Call to a member function getAll() on boolean [/home/flevity/pathfinder/app/main/controller/ccp/universe.php:142]
PHP message: [app/main/controller/api/setup.php:64] Controller\Ccp\Universe->buildSystemsIndex()
PHP message: [index.php:23] Base->run()" while reading response header from upstream, client: 2a02:2698:4c25:5392:9968:afa8:6097:7ef7, server: achtung-wh.tk, request: "POST /api/setup/buildIndex HTTP/2.0", upstream: "fastcgi://unix:/var/run/php/php7.2-fpm.sock:", host: "achtung-wh.tk", referrer: "https://achtung-wh.tk/setup"

after login error:

2018/08/04 17:14:13 [error] 1167#1167: *1 FastCGI sent in stderr: "PHP message: Validation failed: "systemId" = "31000713" [/home/flevity/pathfinder/app/main/model/basicmodel.php:463]
PHP message: [app/main/model/systemmodel.php:242] Model\BasicModel->throwValidationException()
PHP message: [app/main/model/basicmodel.php:311] Model\SystemModel->validate_systemId()
PHP message: [app/main/model/basicmodel.php:184] Model\BasicModel->validateField()
PHP message: [app/lib/magic.php:73] Model\BasicModel->set()
PHP message: [app/lib/magic.php:117] Magic->offsetset()
PHP message: [app/main/model/mapmodel.php:481] Magic->__set()
PHP message: [app/main/controller/api/map.php:993] Model\MapModel->getNewSystem()
PHP message: [app/main/controller/api/map.php:893] Controller\Api\Map->updateMapData()

[Derjyn]

Any updates on this? A bit hard to use this without this functionality.

[alewis-isc]

I'm having exactly this issue as well right now.

[justindh]

Figured out the issue with some trial and error, it looks like the eve_universe cache is struggling to build a system list from scratch. Dunno why, but this fixed it for me.

• In mysql, drop the eve_universe database drop database eve_universe; (your db name may very, checked environment.ini if you are unsure)
• Now create a blank database create database eve_universe;
• Import the eve_universe seeder wiki This step is still required!
• login to /setup and hit the setup tables button on eve-online universe section
• login to /setup and hit the fix Columns/Keys button on eve-online universe section
• clear cache
• login to /setup and hit build system index build button
• login to /setup and hit rest of the index data buttons after each finishes
• profit?

[alewis-isc]

The above worked for me. Good work Sir!

[Flevity]

yeah it is work, tnx

[Tupsi]

tried this a few times now, no dice. still get the wrong counting and the counter never finishes

[lord-carlos]

Running latest /develop branch and still got this problem. (No redis)

Not sure how it should work, but apparently it did not import any systems via ESI. I had to import the .sql file.
Now when I build system index it just stops with server error 500.
/api/setup/buildIndex 500 (Internal Server Error)

2018/11/28 14:02:48 [error] 20441#20441: *21899 FastCGI sent in stderr: "PHP message: session_start(): Failed to decode session object. Session has been destroyed

I'm gonna try it justindh way.

Edit: Ah, updateUniverseSystem and setup cron job are disabled by default.

[exodus4d]

@lord-carlos I was able to reproduce your error. This just happens if "MySQL Sessions" are set in config.ini (which is the default). Installations with" Redis" configured instead of "MySQL Sessions" are not affected.

2018/11/28 14:02:48 [error] 20441#20441: *21899 FastCGI sent in stderr: "PHP message: session_start(): Failed to decode session object. Session has been destroyed

The problem

Building the "system Index" throws a huge amount of temp data into the users Session. The default column type in sessions.data is TEXT which can not handle all and cuts off data ON INSERT... Next time Session data gets accessed fails because of invalid format -> decode failed

Fix

Session DB table now has MEDIUMTEXT instead of TEXT as type for sessions.data column