Slow Page Rendering of Large Pages Due to HTTP POST Limitations
There is no maximum limit stipulated on a page size in Confluence. However, if a page larger then 2Mb is created, a user may experience two problems:
- Rendering of a page is slow. Understandably, rendering of a large page is going to take excessive time. Don't try to put the entire content of a book into one page, split it into pages/chapters. It is easier to manage and fast to render too.
One of the following errors occurs in the logs:Tomcat 5.5.x
java.lang.IllegalStateException: Post too large at org.apache.catalina.connector.Request.parseParameters(Request.java:2361)Tomcat 6.x
Sep 2, 2010 9:33:30 AM org.apache.tomcat.util.http.Parameters processParameters WARNING: Parameters: Invalid chunk ignored.
This error manifests for large pages, on the order of 2 MB or more. Apache Tomcat 5, by default, sets the maximum size of acceptable HTTP POST request to 2MB.
You can reconfigure Tomcat to accept larger requests. This can be done by increasing the allowable limit or just simply disabling this functionality.
- Edit is
Set the Tomcat configuration parameter maxPostSize for the HTTPConnector to a larger value (in bytes) to increase the limit. Setting it to 0 will disable the size check.
If you are using Confluence 5.8.13 (Tomcat 8.0.24 or above), do not change the setting to 0. In Tomcat 8.0.24, the meaning of the value zero for the
maxPostSizehas been changed to mean a limit of zero rather than no limit to align it with
maxSavePostSizeand to be more intuitive. See Tomcat 8.0.24 change log for more details. You must now use maxpostsize="-1" for unlimited size.
See the Tomcat Configuration Reference for more information:
maxPostSize - "The maximum size in bytes of the POST which will be handled by the container FORM URL parameter parsing. The feature can be disabled by setting this attribute
to a value inferior or equal to 0. If not specified, this attribute is set to 2097152 (2 megabytes)."
If using Apache web server, add the same parameter to the AJP connector instead.
- Edit is