Not a bug New category nodes remain invisible until a new forum is made inside of it.
- Thread starter Vekseid
- Start date
This really sounds like permissions didn't rebuild fully. This isn't something I recall any other reports of (and I can't reproduce the behavior mentioned), though it could happen if you actively closed the window while the permission rebuild job was still running. Normally, that would then get listed (within 15 minutes or so) as a message on the control panel "dashboard".
Can you reproduce it consistently? If so, can you confirm the contents of the xf_job table once it's happened?
Can you reproduce it consistently? If so, can you confirm the contents of the xf_job table once it's happened?
Vekseid
Active member
@Mike I just had what I presume is the same issue occur when creating a new forum:
This one's been waiting awhile obviously, was wondering what I did wrong.
Just triggered it again on category creation:
One issue here is that running this permission rebuild job from the admin panel requires more than just the create node permission. My normal 'non-superuser admin' can't run it, because I think it's tied to batch jobs IIRC.
Code:
mysql> select * from xf_job;
+--------+-------------------+----------------------+--------------+----------------+--------------+---------------+
| job_id | unique_key | execute_class | execute_data | manual_execute | trigger_date | last_run_date |
+--------+-------------------+----------------------+--------------+----------------+--------------+---------------+
| 2 | cron | XF\Job\Cron | a:0:{} | 0 | 1544109578 | 1544109518 |
| 9 | xfCollectStats | XF:CollectStats | a:0:{} | 0 | 1547344734 | 1543985471 |
| 43232 | permissionRebuild | XF:PermissionRebuild | a:0:{} | 1 | 1543987106 | NULL |
+--------+-------------------+----------------------+--------------+----------------+--------------+---------------+
3 rows in set (0.00 sec)This one's been waiting awhile obviously, was wondering what I did wrong.
Just triggered it again on category creation:
Code:
mysql> select * from xf_job;
+--------+-------------------+----------------------+--------------+----------------+--------------+---------------+
| job_id | unique_key | execute_class | execute_data | manual_execute | trigger_date | last_run_date |
+--------+-------------------+----------------------+--------------+----------------+--------------+---------------+
| 2 | cron | XF\Job\Cron | a:0:{} | 0 | 1544109944 | 1544109890 |
| 9 | xfCollectStats | XF:CollectStats | a:0:{} | 0 | 1547344734 | 1543985471 |
| 48412 | permissionRebuild | XF:PermissionRebuild | a:0:{} | 1 | 1544109904 | NULL |
+--------+-------------------+----------------------+--------------+----------------+--------------+---------------+
3 rows in set (0.00 sec)One issue here is that running this permission rebuild job from the admin panel requires more than just the create node permission. My normal 'non-superuser admin' can't run it, because I think it's tied to batch jobs IIRC.
These permission rebuild jobs should run regardless of admin permissions. Rebuilding permissions isn't actually something you can run from the "Rebuild caches" admin CP page.
Those jobs should just start running as soon as they are queued up, while the node is saving.
Are you waiting long enough after creating the node for it to kick in and finish? Does it even come up with the "Processing..." screen?
If these jobs do not finish running, you should see this notice on the Admin CP home page:
If you click that, do the jobs finish successfully (and leave the xf_job table)?
Vekseid
Active member
No processing screen. In several cases I never bothered to close the tab for days afterwards.
They don't run until I log in as the superadmin and do this:
This isn't possible for my normal admin users, because it blocks them on permissions.
They don't run until I log in as the superadmin and do this:
This isn't possible for my normal admin users, because it blocks them on permissions.
I've just tested this on 2.0.12 and have not been able to reproduce it.
I created an admin that only had access to manage nodes (no other admin permissions). The job ran as expected after creating a node. I even interrupted that job and waited for the "manual rebuild jobs awaiting completion" message and clicked that -- that kicked the job off as expected.
This all fits with the stock code as
I created an admin that only had access to manage nodes (no other admin permissions). The job ran as expected after creating a node. I even interrupted that job and waited for the "manual rebuild jobs awaiting completion" message and clicked that -- that kicked the job off as expected.
This all fits with the stock code as
tools/run-job is not permission controlled for exactly this reason. At this point, I can only guess that there's an add-on interfering.