Saturday, July 19, 2014

Unstable Aerohive behavior

Recently I had some issues with the entire hive: the APs where restarting randomly and the users were unable to register. immediate hot fix was to restart the units but it was good only for a few days or a week max.

First, diagnostics: the issue we had is completely silent and hidden. No error messages or any other problems show in the logs both on the APs and in the manager. The only warning sign is frequent random restarts of APs: roughly every 3-4 days a few units will restart. Eventually, users that try to authenticate will be refused connection and those who are already authenticated will remain connected.

While working with Aerohive, we found out that it could be due to WIPS (). It is possible that WIPS will force some sort of kernel panic effectively inducing a unit restart. By disabling WIPS, we have extended the restart time and the interval between authentication problems. However the issue was not fixed completely: once a week or a bit more we had to restart everything again.

The solution? We haven't found one. Luckily, a new version came out (6.1r6a). At the moment, it looks like a full upgrade has fixed the issues: no restarts or connection issues for the past week.

No comments:

Post a Comment