Add upgrade documentation
This commit is contained in:
parent
c1611fd3f6
commit
fe17ce2150
34
doc/upgrade.md
Normal file
34
doc/upgrade.md
Normal file
|
@ -0,0 +1,34 @@
|
|||
# Considerations before upgrading Friendica
|
||||
|
||||
* [Home](help)
|
||||
|
||||
## MySQL >= 5.7.4
|
||||
|
||||
Starting from MySQL version 5.7.4, the IGNORE keyword in ALTER TABLE statements is ignored.
|
||||
This prevents automatic table deduplication if a UNIQUE index is added to a Friendica table's structure.
|
||||
If a DB update fails for you while creating a UNIQUE index, make sure to manually deduplicate the table before trying the update again.
|
||||
|
||||
### Manual deduplication
|
||||
|
||||
There are two main ways of doing it, either by manually removing the duplicates or by recreating the table.
|
||||
Manually removing the duplicates is usually faster if they're not too numerous.
|
||||
To manually remove the duplicates, you need to know the UNIQUE index columns available in `database.sql`.
|
||||
|
||||
```SQL
|
||||
SELECT GROUP_CONCAT(id), <index columns>, count(*) as count FROM users
|
||||
GROUP BY <index columns> HAVING count >= 2;
|
||||
|
||||
/* delete or merge duplicate from above query */;
|
||||
```
|
||||
|
||||
If there are too many rows to handle manually, you can create a new table with the same structure as the table with duplicates and insert the existing content with INSERT IGNORE.
|
||||
To recreate the table you need to know the table structure available in `database.sql`.
|
||||
|
||||
```SQL
|
||||
CREATE TABLE <table_name>_new <rest of the CREATE TABLE>;
|
||||
INSERT IGNORE INTO <table_name>_new SELECT * FROM <table_name>;
|
||||
DROP TABLE <table_name>;
|
||||
RENAME TABLE <table_name>_new TO <table_name>;
|
||||
```
|
||||
|
||||
This method is slower overall, but it is better suited for large numbers of duplicates.
|
Loading…
Reference in a new issue