When trying to access Exinda GUI, you're unable to log into the WebUI with the following warning:
Your session is no longer valid
SSH and Console access are still available.
You might also receive a disk full warning and the system may become extremely slow and unresponsive.
show service in CLI displays the majority of services being stopped.
Trying to start DB using
service database start returns An internal error occurred.
In Exinda version 7, authentication towards the appliance is done via the mysql database. If the database is not available then the user will not be able to access the WebUI. Usually, the database is not accessible because
- The user just upgraded their appliance and the database upgrade is still processing in the back end.
- The database gets corrupt, due to a reboot of the device immediately after an upgrade while the MySQL database was still in the process of upgrading.
It is recommended that after an upgrade, even if the graphs are not visible, let the Exinda idle for at least 1-2 hours to let the mysql process upgrade on its own.
It is also possible that the mysql process requires a restart.
- A new process called jboss might require a restart. This process is also in charge of the WebUI.
- CPU utilization increase due to unusual Aborted connections spike
- The browser cookies are not letting the login page load.
- Wait for 20-30 minutes after the upgrade and give your Exinda box time to upgrade the database
- Clear the browser cache and wait for some time (5-10 mins) before trying to access the login page again.
- Force reboot the device from the CLI.
- In the config mode restart the services:
en conf t service jboss restart service datad restart service collectord restart service mysqld restart
Note: If you're prompted for a License key, that means that you need to Activate the Restricted Commands License, open a Support Ticket and ask them to activate it.
- If after 10-15 minutes the issue is still unresolved, it's needed to wipe the database:
pm process mysqld terminate
cd /var/data/monitor rm -rf *
rmcommand is accidentally executed in the wrong location, the damage could be done to the system as there would be no prompting to remove data.
- Afterward, restart the daemons again:
cli en pm process mysqld restart
pm process datad restart
pm process collectord restart
- Wait for 5-10 mins for the processes startup.