Error while updating upm plugin (logExceptions Out of memory)

[jadbaz]

Hello,

I received an error while trying to update upm on my Confluence 6.3.1 server from version 2.2.4 to version 2.2.5.

Below is the tail of atlassian-confluence.log when this happened.

atlassian-confluence.log

2017-08-25 15:46:29,290 INFO [UpmAsynchronousTaskManager:thread-4] [atlassian.plugin.manager.DefaultPluginManager] updatePlugin Updating plugin 'com.atlassian.upm.atlassian-universal-plugin-manager-selfupdate-plugin' from version '2.22.4' to version '2.22.4'
2017-08-25 15:46:29,291 INFO [UpmAsynchronousTaskManager:thread-4] [atlassian.plugin.manager.DefaultPluginManager] broadcastPluginDisabling Disabling com.atlassian.upm.atlassian-universal-plugin-manager-selfupdate-plugin
2017-08-25 15:46:33,846 ERROR [pool-919-thread-1] [engine.jdbc.spi.SqlExceptionHelper] logExceptions Out of memory (Needed 8112696 bytes)
 -- url: /rest/plugins/self-update/1.0/ | referer: http://wiki.activeidentity.com/plugins/servlet/upm | traceId: bc47e40aeec0cd3c | userName: X
2017-08-25 15:46:33,846 WARN [pool-919-thread-1] [confluence.impl.hibernate.ConfluenceHibernateTransactionManager] doRollback Performing rollback. Transactions:
 ->[PluginReadWriteTx]: PROPAGATION_REQUIRED,ISOLATION_DEFAULT (Session #239037332)
 -- url: /rest/plugins/self-update/1.0/ | referer: http://wiki.activeidentity.com/plugins/servlet/upm | traceId: bc47e40aeec0cd3c | userName: X

A few more questions:

1. Isn't it odd that the log is saying "from version '2.22.4' to version '2.22.4'"
2. Does this have to do with inno_db_log_file_size? I'm 100% sure I fixed that. show variables on my RDS instance returns: innodb_log_file_size=26214400

Here is my Confluence server information

[AnnWorley · Accepted solution]

I searched our support cases for the message that the UPM was updating "from version '2.22.4' to version '2.22.4'" and was not able to find any precedent for that message. It is odd, as you pointed out.  It is an INFO message so I am more concerned with the ERROR entry later in the log. It does look like the system ran out of memory as suggested by Nic on Friday:

Error updating upm ([engine.jdbc.spi.SqlExceptionHelper] logExceptions Out of memory)

While you are restarting the system to pick up the memory changes, please also clear the plugins cache as described in: How to clear Confluence plugins cache

I look forward to hearing the result.

Please update the other thread instead of this one, so we can keep Nic in the loop.