Bulk auto-copying times out with large portfolios and number of people

Bug #1862892 reported by Kristina Hoeppner
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
Mahara
Confirmed
High
Unassigned

Bug Description

We can auto-copy portfolios in a few ways:

- Site portfolios to all new registrations
- Institution portfolios to all new institution members
- Group portfolios to all existing group members

We've had reports that this copying can time out, e.g. a group with 70 members and a portfolio collection with 7 pages timed out.

Another report was on the institution level of copying multiple collections into more than 100 newly created accounts.

The funny thing is that if you let Mahara be on its own (even navigating away from the page after a timeout), it will eventually finish the copying of the portfolios. However, the interface doesn't yet support that gracefully.

For the group portfolio copying the following happened (clean URLs turned on):

1. Copying was kicked off.
2. It looked to be finished / timed out.
3. One account was checked and the portfolio was not there.
4. The copying was kicked off again.
5. Some accounts ended up with the collections twice and some page in the collection had the exact same title, which suggests that the database processes overlapped and the database didn't have the first page name written yet when the second came but then still continued with the first name and thus had two pages of the same name. That rendered the collections unusable and didn't display the required pages.

I think it would be good to have a queue process that can run in the background so that the front-end can be continued to be used and the group admin, institution admin or site admin be informed when the bulk copying has finished. This is usually only a problem when people are set up via CSV file or any other process where lots of accounts are created in bulk (for site and institution copying) or in a group.

We may also need to find out the threshold first for when problems occur if we wanted to limit the notifications or queue to a certain size or if we say to have a low threshold for institution and site portfolio copying and always employ the queue in groups.

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.