User:Victor Grousset - tuxayo:Draft:Real world upgrade feedback
Jump to navigation
Jump to search
Aim of the page
- time of updatedatabase.pl
- disk space
- potential failure causes
Performance and times
Case 1
- from 3.22 to 17.11
- 150k biblios, 130k authorities
- ~10000 active patrons
- mechanical HDD
- CPU: Intel from 2015 4 cores@2.6GHz
- container environment (LXD)
- 29 min of updatedatabase.pl
- 86 min of misc/maintenance/auth_show_hidden_data.pl -c
Case 2
- from 3.22 to 17.11
- Diskpace: 1GiB/100k biblio records due to bug 17196: Move marcxml out of the biblioitems table. During the 16.11 to 17.05 part.
- virtualized environment (VMware)
Case 3
- from 3.22 to 17.11
- 300k authorities
- 74 min of misc/maintenance/auth_show_hidden_data.pl -c
- mechanical HDD
- virtualized environment (VMware)
Case 3
- from 16.11.15 to 17.11.06
- 65k biblios, 50k authorities
- no active patrons, only on site reading
- container environment
- 42 min of misc/maintenance/auth_show_hidden_data.pl -c
Potential failure causes
16.11 → 17.05
Duplicates in deletedbiblioitems
Will cause one instruction of the "16.12.00.004" update fail. (INSERT INTO deletedbiblio_metadata)