When running Confluence on AWS Aurora i am getting this error
Your innodb_log_file_size of 50,331,648 is too small. You should increase innodb_log_file_size to 256M.
Aurora does not allow to change this parameter.
First, is it a real problem? What could be the consequences?
Second, is there any workaround, e.g. connection string suffix or so?
Yes, you can lose writes to the database and risk crashes.
You need to increase the log file size. Sorry, there's no other option.
Again, you need to increase the log file size.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Followup on this question, has anybody actually had a problem using confluence on Aurora? The reason I ask this is that since the db engine behind the scenes isn't actually innodb, but a customized version of innodb for Aurora, there may not actually be an issue with the "size" of the innodb log, as it may not behave the same way. This may simply be a placeholder value that Aurora is reporting to applications that ask, and thus why the value isn't something you can change. I don't know if this is the case, but anybody reporting if this actually IS an issue, vs. one confluence simply reports as an issue would be nice.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
I tried it with Atlassian's #1 implementer and it couldn't be done. Confluence had loading issues and could not bring up add-ons. I would recommend going with Postgress as recommend above, that's what I ended up doing.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
I installed Confluence with Aurora for MySQL, and upped the max packet size to 256, and was able to install the draw.io addon, and everything is working fine, including after a restart. This may be an issue that has been resolved at some point, or I just haven't stumbled on an addon that is large enough to cause a problem.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hello!
I am facing this same issue right now. Can you throw some light on how you were able to increase the max packet size?
Thanks!
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
The database parameters allows you to adjust the max packet size, but not the innodb log size. So far, nothing has had an issue with this.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
FWIW, this is Amazon Support's answer to this question:
The parameter "innodb_log_file_size" is not applicable for Aurora MySQL and doesn't need to be tuned in Aurora. In MySQL, increasing this variable might be helpful to increase the performance of the write operations but it is not valid for Aurora as it uses different storage to handle all write operations effectively. Kindly note that RDS Aurora does not handle logs same way as RDS MySQL instances does, therefore this parameter is inconsequential in RDS Aurora MySQL. This is the reason it is not exposed in the parameter group of your Aurora DB instance. Your tool may detect this but RDS Aurora does not use it. So, you can kindly ignore the message. Having said that, if you are facing any issues regarding this, please let us know. We'll be happy to help you. Similarly, several Aurora parameters are optimized for better performance and cannot be changed in Aurora. You can find a list of the variables available in Aurora RDS in the documentation below:
I hope this helps. I am building a Confluence server on Aurora as well.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.