blob: 04b409a9306eac8370f76805ac0b607f5c97d885 (
plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
|
# frozen_string_literal: true
module Gitlab
module Database
module Reindexing
# Number of indexes to reindex per invocation
DEFAULT_INDEXES_PER_INVOCATION = 2
SUPPORTED_TYPES = %w(btree gist).freeze
# When dropping an index, we acquire a SHARE UPDATE EXCLUSIVE lock,
# which only conflicts with DDL and vacuum. We therefore execute this with a rather
# high lock timeout and a long pause in between retries. This is an alternative to
# setting a high statement timeout, which would lead to a long running query with effects
# on e.g. vacuum.
REMOVE_INDEX_RETRY_CONFIG = [[1.minute, 9.minutes]] * 30
# candidate_indexes: Array of Gitlab::Database::PostgresIndex
def self.perform(candidate_indexes, how_many: DEFAULT_INDEXES_PER_INVOCATION)
IndexSelection.new(candidate_indexes).take(how_many).each do |index|
Coordinator.new(index).perform
end
end
def self.cleanup_leftovers!
PostgresIndex.reindexing_leftovers.each do |index|
Gitlab::AppLogger.info("Removing index #{index.identifier} which is a leftover, temporary index from previous reindexing activity")
retries = Gitlab::Database::WithLockRetriesOutsideTransaction.new(
timing_configuration: REMOVE_INDEX_RETRY_CONFIG,
klass: self.class,
logger: Gitlab::AppLogger
)
retries.run(raise_on_exhaustion: false) do
ApplicationRecord.connection.tap do |conn|
conn.execute("DROP INDEX CONCURRENTLY IF EXISTS #{conn.quote_table_name(index.schema)}.#{conn.quote_table_name(index.name)}")
end
end
end
end
end
end
end
|