Solr failure notifications during 06-22-2019 update
Starting this morning every 10 minutes I'm getting the following email exception from cpanel. I'm on version 78.0.30. It stopped at noon but now again I'm receiving the exception every 10 minutes. I did not run any update or any maintenance manually. Any suggestion?
Cpanel::Exception::HTTP::Network/(XID wky2qv) The system failed to send an HTTP "GET" request to "http://(here is ID):__HIDDEN__@127.0.0.1:8984/solr/dovecot/update?commit=true" because of an error: Could not connect to '127.0.0.1:8984': Connection timed out
at /usr/local/cpanel/Cpanel/HTTP/Client.pm line 107.
Cpanel::Dovecot::Solr::Utils::call_commit() called at /usr/local/cpanel/3rdparty/scripts/cpanel_dovecot_solr_commit line 40
scripts::cpanel_dovecot_solr_commit::run("scripts::cpanel_dovecot_solr_commit") called at /usr/local/cpanel/3rdparty/scripts/cpanel_dovecot_solr_commit line 35
-
Adding, before I started getting the exceptions I got this update from the daily yum. /etc/cron.daily/0yum-daily.cron: The existing index (version 6) is too old for the current Solr version. The system will rebuild the index in the background. Renamed index directory to: /home/cpanelsolr/server/solr/dovecot/data/index.1561281014 stat(/etc/cpanel-dovecot-solrdisable): No such file or directory - proceeding as though the file is absent. Warning: cpanel-dovecot-solr.service changed on disk. Run 'systemctl daemon-reload' to reload units. The system is deleting this directory in the background. (PID 16830) Restarting Solr... Waiting for "cpanel_dovecot_solr" to start ""waiting for "cpanel_dovecot_solr" to initialize "finished. cpanel_dovecot_solr started successfully. The system is rebuilding user indexes in the background. (PID 17072) Id: TQ:TaskQueue:25339 ================================================================================ Package Arch Version Repository Size ================================================================================ Updating: cpanel-analytics noarch 1.3.3-5.2 cpanel-plugins 25 k cpanel-dovecot-solr noarch 8.1.1-4.1 cpanel-plugins 24 M cpanel-dovecot-solr-contrib noarch 8.1.1-4.1 cpanel-plugins 69 M cpanel-dovecot-solr-contrib-extraction noarch 8.1.1-4.1 cpanel-plugins 38 M cpanel-dovecot-solr-docs noarch 8.1.1-4.1 cpanel-plugins 4.9 M cpanel-dovecot-solr-server noarch 8.1.1-4.1 cpanel-plugins 65 M cpanel-letsencrypt noarch 2.25-2.1 cpanel-plugins 37 k Transaction Summary ================================================================================ Upgrade 7 Packages The updates were successfully applied 0 -
Recently we released a patch to Dovecot Solr which ran a function that did not properly read older configuration settings for Dovecot, leading to your SSL/TLS settings being set to default, which is TLSv1.2. We are tracking this issue internally with case ID CPANEL-28089. The workaround is rather simple, here are the steps: 1. Login as root to WHM 2. Enter WHM >> Home >> Service Configuration >> Mailserver Configuration 3. Change "SSL Minimum Protocol" to the correct value. Most servers affected by this were running TLSv1, for the sake of Windows 7 TLS support. If you have not saved any settings in Mailserver Configuration yet and are unsure of what it was previously set as, you can find the value by running this command: grep -P 'ssl_(min_)?protocol' /etc/dovecot/dovecot.conf
And the example output:grep -P 'ssl_(min_)?protocol' /etc/dovecot/dovecot.conf ssl_min_protocol = TLSv10 -
Hello @Matias Lopez, New Solr RPMS (part of the Full Text Search Indexing for IMAP powered by Apache Solr" plugin offered in WHM >> Manage Plugins) were published on June 22nd, 2019. This particular update automatically rebuilt the Apache Solr" index files due to a format change in the new version. This lead to the error messages like the ones reported in your initial post on some systems due to the service not restarting quickly enough upon the update. This should only have happened once during the update process, and the cpanel_dovecot_solr should have started successfully after the update process completed. Let us know if you encounter any additional issues. Thank you. 0
Please sign in to leave a comment.
Comments
3 comments