How to determine uptime intervals from gateway logs?

[Vladimir]

Does anyone know if there is a way to extract the history of boot and shutdown dates and times from R80.10 vSEC instances?

[PhoneBoy]

Here's what I see in my vSEC instances in AWS.

These gateways are set up in an autoscaling group and are created/destroyed using an Ansible playbook.

Your mileage may vary. 

[Vladimir]

Thanks,

Unfortunately, this will not work for me: I have number of vSECs non-AG that are periodically powered-up or down, but the management server itself suffered failure and was restored from the snapshot.

Hence this headache. Do you know if there are any events pertaining to power-up/down logged on the gateways themselves?

[PhoneBoy]

Every Unix-like system will have tell-tale signs of a reboot in /var/log/messages.

I believe by default we keep 5 days worth of messages.

And there's obviously the "uptime" command on each gateway which will tell you when it was last rebooted.

[PhoneBoy]

I think what you are after might be in $FWDIR/log/fwd.elg, though... 

[Vladimir]

native will not work, as I am looking for about a month and a half of halt/powerup events.

Same issue is with "uptime", as it only shows the current value.

[Vladimir]

fwd.elg does not contain the timestamps.

bummer...

[Hugo_vd_Kooij]

Unfortunatly the default values may not work to well. I recommend to:

1. Increase the number of backup files and the size of each Messages file. sk36798
2. If you get bogus input like missing RAID disks. Find the cause and make sure it is no longer send to the messages file when it is not needed.

And I strongly recommend to send your syslog to a central log server.

[Vladimir]

Thank you.

This is a one-off situation. I've build a bunch of AWS scenarios intended to be a proof of concept, but am now in a position to bill the client for the time spend on it.

Under normal circumstances, (i.e. in production), I'll be using an  AWS Cloud Trail to have this info readily available. But since it was not configured from the get-go, I was trying to determine if there are CP specific logs with this data.

Alas, I'll have to estimate time spent on the project without supporting documentation, which I really dislike doing.