BlueWiki maintenance
Debug of a running application
Change is instant and all errors could see everyone with access to the url!!!
cat > settings.d/9999-debug.php<<EOF
<?php
$wgShowSQLErrors = true;
$wgDebugDumpSql = true;
$wgShowExceptionDetails = true;
$wgShowDBErrorBacktrace = true;
$GLOBALS['wgDebugLogGroups']['OAuth'] = '/data/php-fpm/logs/oauth-debug.log';
EOF
Update
Image in our registry
When bluespice build a new image they add a new tag. So get this new tag and:
- go to git.easy.cz/bluespice/bluespice
- edit file build-MAJOR.MINOR/VERSION and increase the version
- edit file build-MAJOR.MINOR/Dockerfile and update source image adress
- commit changes with note “UPDATE - $NEW_VERSION”
Wait till the pipeline finish…
Deployment
Simple redeploy application with newer docker image tags (Should be implemented in ESko)
Monitoring
Works like monitoring of the common EP/ER application. In short:
When the application has been deployed from ESko as an client application
- application has been registered to icinga monitoring and /sys/monitoring.json endpoint will be monitored (we made this endpoint as a static file in /data/bluespice/sys/monitoring.json)
We create this file in operator container by init.sh script
Backup
-
Databases are backup as usual, because our backuping system works well with all kind of mysql like databases.
-
Files should be backuped easily. The directory with static files is /data/bluespice/w/images/
-
Elasticsearch could be backuped by snapshoting, but we don’t have to backup it, because it could be regenerated after application restart
Restore
-
Database is easy, because we already have a mechanism for this implemented in our infrastructure.
-
Files is a bit more complicated, because it does not have the same structure as our EP, in initial phase we will skip it and will do it manualy
-
Elasticsearch we doesn’t backup it.