You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Notice that using memory-mapped I/O may not help performance in the highly concurrent environment of a busy web server.
151
+
150
152
= I sometimes get timeout errors from SQLite. How can I fix them? =
151
153
152
154
Some sites occasionally generate error messages looking like this one:
153
155
154
156
`Unable to execute statement: database is locked in /var/www/wp-content/object-cache.php:1234`
155
157
156
-
This can happen if your server places your WordPress files on network-attached storage (that is, on a network drive). To solve this, store your cached data on a locally attached drive. See the question about storing your data in a more secure place.
158
+
This can happen if your server places your WordPress files on network-attached storage (that is, on a network drive). To solve this, store your cached data on a locally attached drive. See the question about storing your data in a more secure place. It also can happen in a very busy site.
157
159
158
160
= Why do I get errors when I use WP-CLI to administer my site? =
159
161
@@ -191,6 +193,8 @@ Please look for more questions and answers [here](https://www.plumislandmedia.ne
191
193
192
194
* Clean up in chunks in an attempt to reduce contention delays and timeouts.
193
195
* Do PRAGMA wal_checkpoint(RESTART) when cleaning up, and also occasionally, to prevent the write-ahead log from growing without bound on busy systems.
196
+
* Retry three times if cache updates time out.
197
+
* Increase default cache size to 16MiB for new users.
194
198
195
199
= 1.3.5 =
196
200
@@ -210,6 +214,6 @@ Please look for more questions and answers [here](https://www.plumislandmedia.ne
210
214
211
215
== Upgrade Notice ==
212
216
213
-
This release attempts to reduce cache timeouts by doing cleanup operations in chunks. It also does PRAGMA wal_checkpoint(RESTART) when cleaning up, and also occasionally, to prevent the write-ahead log from growing without bound on busy systems.
217
+
This release attempts to reduce cache timeouts by doing cleanup operations in chunks, and by retrying timed-out cache update operations. It also does PRAGMA wal_checkpoint(RESTART) when cleaning up, and also occasionally, to prevent the write-ahead log from growing without bound on busy systems.
214
218
215
219
Thanks, dear users, especially @bourgesloic, @spacedmonkey, @spaceling and @ss88_uk, for letting me know about errors you found, and for your patience as I figure this out. All remaining errors are solely the responsibility of the author.
0 commit comments