2023-02-01 23:12:33 |
Galen Charlton |
bug |
|
|
added bug |
2023-02-01 23:12:45 |
Galen Charlton |
tags |
|
circ-holds |
|
2023-02-02 13:50:34 |
Jeremy Murray |
bug |
|
|
added subscriber Jeremy Murray |
2023-02-02 16:08:17 |
Galen Charlton |
description |
Current, one part of opportunistic hold capture is retrieving a set of up to 100 nearest candidate holds from open-ils.storage.action.hold_request.nearest_hold, then going through the list and looking for the first one that passes the permit test.
However, two problems recently seen with this are:
- For a popular title, it is possible to get a set of 100 candidate holds where _none_ of them pass the permit check
- The additional individual permit checks for each hold can take ~10 seconds in that worst case
I've experimented with putting the invocation of action.hold_retarget_permit_test() directly in the nearest_hold query and... it just might help. |
Currently, one part of opportunistic hold capture is retrieving a set of up to 100 nearest candidate holds from open-ils.storage.action.hold_request.nearest_hold, then going through the list and looking for the first one that passes the permit test.
However, two problems recently seen with this are:
- For a popular title, it is possible to get a set of 100 candidate holds where _none_ of them pass the permit check
- The additional individual permit checks for each hold can take ~10 seconds in that worst case
I've experimented with putting the invocation of action.hold_retarget_permit_test() directly in the nearest_hold query and... it just might help. |
|
2023-02-02 16:09:24 |
Galen Charlton |
description |
Currently, one part of opportunistic hold capture is retrieving a set of up to 100 nearest candidate holds from open-ils.storage.action.hold_request.nearest_hold, then going through the list and looking for the first one that passes the permit test.
However, two problems recently seen with this are:
- For a popular title, it is possible to get a set of 100 candidate holds where _none_ of them pass the permit check
- The additional individual permit checks for each hold can take ~10 seconds in that worst case
I've experimented with putting the invocation of action.hold_retarget_permit_test() directly in the nearest_hold query and... it just might help. |
Currently, one part of opportunistic hold capture is retrieving a set of up to 100 nearest candidate holds from open-ils.storage.action.hold_request.nearest_hold, then going through the list and looking for the first one that passes the permit test.
However, two problems recently seen with this are:
- For a popular title in a consortium where, generally speaking, not all copies can fill all holds on the title, it is possible to get a set of 100 candidate holds where _none_ of them pass the permit check
- Running the additional individual permit checks for each hold can take ~10 seconds in that worst case
I've experimented with putting the invocation of action.hold_retarget_permit_test() directly in the nearest_hold query and... it just might help. |
|
2023-02-02 16:10:01 |
Elizabeth Davis |
bug |
|
|
added subscriber Elizabeth Davis |
2023-02-02 17:36:09 |
Ruth Frasur Davis |
bug |
|
|
added subscriber Ruth Frasur |
2023-02-02 17:54:05 |
Christine Morgan |
bug |
|
|
added subscriber Christine Morgan |
2023-02-02 18:00:04 |
Joan Kranich |
bug |
|
|
added subscriber Joan Kranich |
2023-02-02 18:29:41 |
Lindsay Stratton |
bug |
|
|
added subscriber Lindsay Stratton |
2023-02-02 21:06:24 |
Mary Beth Long |
bug |
|
|
added subscriber Mary Beth Long |