My recent CSV imports do not work, the associated jobs are marked as “Starting” and nothing is listed under “Past imports” but I think they are blocked by the Batch Delete (Job 1). The Batch Delete was unsuccessful because I had mistakenly made a change in AbstractEntityAdapter.php (my bad), but is there a way now to undo or stop this job, other than re-installing the whole Omeka?
This actually reminds me very much of the problem detailed here for Omeka 2, is the procedure to follow the same?
I have a few job which have been exited without returning. Their status is still “In Progress”, but I can start new jobs.
When you click on a Job Id, you get the job page and there, in the top-right corner there is an “Attempt Stop” button. You can press it but I am pretty sure that it will just turn the status on “Stopping”, nothing else.
The only case getting “Starting” status I have met, this have been with a server configuration issue. Is the package PHP-CLI correctly installed on your server? Do you run PHP in safe mode?
I tried to delete the tables of the website and build it anew but I get the same status “Starting” for my jobs (CSV imports).
Pardon my ignorance but searching for Safe mode for PHP, one finds it has been DEPRECATED as of PHP 5.3.0 and REMOVED as of PHP 5.4.0.
Also it seems that PHP-CLI is not correctly installed on the server, we are looking at this possibility.
Why it is important? Can you explain a bit what is behind this?
The php cli is used to process all background jobs, so you can’t csv-import anything without it. Normally, it is autodetected by Omeka/Omeka S, but if it fails, you have to set it in the config of Omeka.
The php cli (command line interface for php) can be same than the program (php) used by the web server (apache or nginx). You just have to set its path in the config.You can install multiple php on the same machine, but it is not required.
Thanks for the help. Setting the php_cgi path fixed part of my problem. When the error went away I was stuck with the Starting job status with nothing more. Running from command line caused a Fatal error with LDAP. A bit more about that here.
I take the liberty to hijacked this thread because the title refers exactly to an issue I have. I could open another thread if needed but I believe it is more useful for the future searches this way.
I have a solr indexing job which failed. Solr core was restarting during the indexing I guess. This job was in statut “en cours” in Omeka S job management page. I stopped it and it is now in status
“Arrêt en cours”.
Since this, each time I launch a solr indexing process, even when I force the job, I have the following message:
2019-03-14T07:37:04+00:00 INFO (6): Search index #1 ("solr"): start of indexing
2019-03-14T07:37:04+00:00 ERR (3): Search index #1 ("solr"): end of indexing: there are already 1 other jobs "Search Index" and the current one is not forced.
How can I definitively get rid of this stucked job which blocks future jobs?