Forum Discussion

steve_laffan_11's avatar
steve_laffan_11
Icon for Cirrus rankCirrus
Jun 23, 2015

AWS F5 11.6.0.4.0.420-HF4 Failing to reload

Hi

 

Been working lots on F5 virtual appliances inside AWS and have seen many occurrences of the virtual appliance not coming back up after a reboot....a second or third reboot occasionally resolves this sometimes, but other times I have had to recreate the appliance again. I am running the hourly license on all of them and using a mix of 25Mbps through to 1Gb on good, better and best feature sets but encounter the same failed to restart issue.

 

I can't see to find much about the issue and have spoken to F5 who have advised to raise a case...difficult when we are just using hourly licenses only. On a failed startup, the GUI shows 'Configuration Utility Restarting' and never recovers. At this time, I SSH via root to appliance and can see in /var/log/ltm the following recurring messages:

 

Jun 23 15:50:09 ip-10-128-5-93 err mcpd[4734]: 010717ff:3: [Licensing]: Failure in establishing instance identity. Jun 23 15:50:09 ip-10-128-5-93 err mcpd[4734]: 01070734:3: Configuration error: MCPProcessor::check_initialization:

 

I suspect this is down to either the hourly license or the code version...I keep testing this on the only other F5 version available within AWS which is 'BYOL Hotfix-BIGIP-11.4.1-HF8-681.73-ENG' and this seems to reboot perfectly every time...albeit that it does not have a license applied. Any assistance would be appreciated

 

Cheers Steve

 

APM
cloud
security
TMOS
TMSH
virtualization

5 Replies

Sort By
  • steve_laffan_11's avatar
    steve_laffan_11
    Icon for Cirrus rankCirrus
    Jun 23, 2015
    So after a little more testing, this issue seem to only occur when using the per hour licensing model. I have loaded, reloaded and shutdown instances with BYOL licenses and they work perfectly..... Suspect this is going to be an F5 tech resolution only !!
  • steve_laffan_11's avatar
    steve_laffan_11
    Icon for Cirrus rankCirrus
    Jun 25, 2015
    Thanks Daniel. Will be very interested in seeing what you find as I can reproduce the issue almost at will and it is impacting a number of AWS F5 services I am working with. Steve
  • Daniel_Epperson's avatar
    Daniel_Epperson
    Icon for Employee rankEmployee
    Jun 25, 2015
    Hi Steve, I have a few questions for you. 1. What process are you using to restart the instances? 2. Are you trying to make any changes to the instance since it was deployed? 3. Can you provide me the exact AMI id and Availability Zone you are using? 4. From a recovered unit, can you collect a "quick view" (qkview from cli)? In your case it appears that something at the "hardware" layer is changing. We want the hardware to look basically exactly the way it did when the instance first started up. If you need to change to a bigger instance, deploy a new instance with the new instance type and migrate the configuration. In my lab, I wasn't able to replicate your issue using F5 Networks Hourly Hotfix-BIGIP-11.6.0.4.0.420-HF4 - Good 25Mbps (ami-297dc634) using a variety of methods. I tried rebooting the unit via SSH, sending reboot command from AWS Console Instance State menu and manually stopping then starting the instance. In all cases my instance came up licensed okay. Maybe your issue is AZ or process specific. Using this this CloudFormation Template allowed me to very quickly deploy a new BIG-IP and VPC, you may find it useful as well. https://devcentral.f5.com/s/articles/aws-cft-for-hourly-big-ip-116-in-ec2 If you are able to, please upload a qkview file from one of your recovered instances so that I can look at the log files in detail. If you can access ihealth.f5.com then please upload the qkview directly there. If not, can you provide me your email address so that I can arrange a place for you to upload the files? Thanks! --Danny
  • Daniel_Epperson's avatar
    Daniel_Epperson
    Icon for Employee rankEmployee
    Sep 24, 2015

    F5 resolved this issue as bug ID532145 in engineering hotfix 11.6.0.5.50.429.50. Please open a support case (preferred) or PM me if you need this fix.