The identification page crashes, and prevents authentication

Question

Hello,

From time to time (about every 96 hours), and on as many computers as web browsers, the H-MDM identification panel freezes. We enter the Admin username, and the password, and nothing happens. The "authenticate" button does not respond. 

Restarting the server or TomCat9 services is mandatory.

No error log from tomcat9, nor from ubuntu. It seems to be directly in the code (js?) which prevents identification. the error also occurs in localhost on the machine.

I wanted to redo the test on a VM, with another domain name. The frequency of blogging is less, but it also appears. I left the VM running for 1 week, and on Sunday the "authenticate" button was not working. As much on my phone in 4G connection, as much on my computers in WIFI.

What do you think ?

Thank you 

Comments

Can you please open the Chrome developer console and look for the JavaScript errors? Does refreshing the page help? Note that you must use https for login: plain HTTP doesn't work.

---

I actually use https connection.
As soon as the crash occurs again, I look at the chrome console.
Refreshing the web page does not change anything, as much on Chrome, as on Firefox, as on Safari on a mobile internet connection... So with a different IP...
After restarting tomcat9, the connection works again...
To be continued on the next crash, I don't think I will have any logs before 96 hours
thank's

Answer 1

Sometimes Headwind MDM web application gets a kind of an "Application Not Responding" issue. Unfortunately the error cause has not yet been found, but the issue can be diagnosed by 100% CPU usage.

Take a look in the "cpu_monitor.sh" script inside the installer archive. Review it, change the notification emails, and set it to the crontab to be executed, say, each 10 minutes. This should automatically restart Tomcat and fix your issue.

To add the script to crontab, copy it in some fixed directory, for example, /opt/hmdm, then edit the crontab by a command (as root):

crontab -e

Then add the following line:

*/10 * * * * /opt/hmdm/cpu_monitor.sh

Make sure the rights to execution are granted to this script. If it has no execution permissions, add them by the command:

chmod +x /opt/hmdm/cpu_monitor.sh

Comments

Excellent idea. Do you just have a procedure to add it to the crontab?
When the WEB application crashed, the machine's processor was at 20%. But, indeed, why not parameterize your script.
If you ever have time, why not do a detailed procedure in this ticket, so that it could be useful to others! :)