Ticket #894 (assigned Issue)
why are there missing indices in the PG database?
| Reported by: | charlotte | Owned by: | charlotte |
|---|---|---|---|
| Priority: | minor | Milestone: | V1.2 Questionnaire release |
| Component: | WP6 - CMIP5 Questionnaire | Version: | |
| Keywords: | Cc: | gerry | |
| Requirement: | |||
Description
There are about 1500 unused primary key indices in the postgres database. Why?
Change History
comment:2 Changed 3 years ago by charlotte
- Status changed from new to assigned
The missing indices did not appear when the questionnaire was redeployed. Instead they appeared sometime between the 13th and 21st of October. The number of missing indices remains constant - no further gaps have appeared in the protoq_component table.
| Date | Primary key | protoq_component id | difference |
| 10th Oct | 1-140 | 1-140 | 0 |
| 11th Oct | 1-140 | 1-140 | 0 |
| 12th Oct | 1-840 | 1-840 | 0 |
| 13th Oct | 1-840 | 1-840 | 0 |
| 21st Oct | 1-840 | 1-840 | 0 |
| 841-1261 | 2043-2463 | 1202 | |
| 22nd Oct | 1-840 | 1-840 | 0 |
| 841-1611 | 2043-2813 | 1202 | |
| 9th Nov | 1-840 | 1-840 | 0 |
| 841-3155 | 2043-4357 | 1202 | |
| 10th Nov | 1-840 | 1-840 | 0 |
| 841-3225 | 2043-4427 | 1202 | |
comment:3 Changed 3 years ago by charlotte
Hi Charlotte,
the backup on the 12th appears to match yours (i.e. goes up to 840 with no missing blocks). I think this is therefore not a case where data got deleted, rather something strange happened in the incrementing of the id.
Gerry
The 17th November backup has missing indices - Gerry's local backup made on the 13th has missing indices too. I thought the DB was redeployed on the 12th.
| 10th October : Deployed official release version 1.0 |
| 12th October : Implemented validation fixes. |
| 21st October : migrated to Python 2.6 |
| 9th November: small changes for Mark E |
| 10th November : small changes made for Mark E |
How come the missing indices appear on the 13th
comment:4 Changed 3 years ago by charlotte
| Date | Primary Key | protoq_component id | difference |
| 14th Oct | 1-840 | 1-840 | 0 |
| 841-980 | 2043-2182 | 1202 |
So the missing indices occur when model "JustATest" is created in the "2. Test" centre (number 39)
is it possible that an offset gets introduced when we redeploy the questionnaire? Would like to get to the bottom of this before we implement submit to CMIP5.
I have requested before and after backups for each redeployment of the questionnaire.