CPage exception in CMain/renderHTTP

[lcasey001]

Running latest 2.7.5 and we are getting the following error multiple times a day. At times doing a db restart help others only a complete server restart.

Doing a sacli stop/start/DBReset and even a Reset does not get it back to a working state. After a reboot it works for a while then back to the same error.

When this occurs no new connection to the UI or Admin page is allowed.

2019-10-15T16:14:08-0400 [stdout#info] [WEB] OUT: '2019-10-15T16:14:08-0400 [stdout#info] u"CPage exception in CMain/renderHTTP: sqlalchemy.exc.StatementError: (sqlalchemy.exc.InvalidRequestError) Can\'t reconnect until invalid transaction is rolled back\\n[SQL: SELECT config.profile_id AS config_profile_id, config.name AS config_name, config.value AS config_value \\nFROM config INNER JOIN profile ON profile.id = config.profile_id \\nWHERE config.name IN (%s) AND profile.name = %s]\\n[parameters: [{}]] (9000)"'

FYI, most users are now using v3 clients.

[novaflash]

This is a problem that occurs very rarely and we are currently investigating it and hope to have it resolved in 2.8. Sorry for the inconvenience. We have no solution at this time.

[lcasey001]

Any idea what we can do in the meantime? This is definitely not a rare case for us and happens constantly ?

[novaflash]

It's still being investigated, so, no, I don't.

Just some random suggestion: you say you upgraded. Maybe your OS is old and could use a reinstall, and then you could choose to migrate your data to the new OS?

[lcasey001]

We have actually reverted back to 2.7.4 and so far the issue has not occurred. However the configs from 2.7.5 no longer work on 2.7.4. That is another issue which we submitted a ticket for.

[lcasey001]

For transparency we decided to use the latest 2.7.5 appliance versus doing an in place upgrade and have so far only seen this issue once in a 24hr period. A service restart got it working again.

[novaflash]

Are you using a MySQL database backend or the default SQLite databases? Are you using clustering feature or just stand alone? If you use MySQL, where is this MySQL server and what type is it?

[lcasey001]

We are using an AWS RDS Aurora instance now. The previous 2.7.4 inplace upgrade to 2.7.5 was using an AWS RDS MYSQL instance. The 2.7.5 instance is a fresh installation.

[novaflash]

That seems to fit the information we have here as well, that it happens when using AWS RDS. We'll fix it and release AS 2.8 when that's done.

[lcasey001]

Is there an eta or a issues tracker for the 2.8 release ?

[novaflash]

Sorry, not publicly visible. And it will be done before the end of the year is the best estimate I can give you. If you need it stable now, then use SQLite databases - the default that comes with AS.

[skouide]

Any idea on how when it will be fixed?

We are facing this issue on a regular basis, this is unacceptable when our corporate users can't even log in.

PS: we use OpenVPN 2.7.5 with RDS as well.

[novaflash]

When AS 2.8 is released it will be fixed. Don't pin me on this, this is just a wild estimate, but somewhere beginning of december it should be out. I can not be any more specific than that.

[evs]

I am having the exact same problem. The only difference is that I am not using AWS RDS, I have a MySQL instance running in the same server I have the OpenVPN server.

[gposton]

I have the same problem. I'm on v2.7.5 connecting to RDS Aurora MySQL. Any word on the ETA for the fix and/or v2.8 release?

[skouide]

I can't believe we are still waiting for a fix, last release was in August...