Go Back
1 Vote

LMU (License Manager Utility) Should Be Enhanced To Support Validation Switches That Return Non-Zero Exit Codes


LMU (License Manager Utility) https://www.synergex.com/docs/index.htm#icg/icgChap2Lmuutility.htm should be enhanced to support the following validation functions:

  • -v
  • -vp
  • -vb
  • -vb SERVERNAME
Which should do the following. All commands should return a 0 exit code in case of success and a non-zero exit code in case of failure

-v / Verify License Servers
This should verify ALL License Servers are running (both primary and backup servers) accessible. Returning a non-zero exit code if any are not working.

-vp / Verify Primary
This should verify that the Primary License Server is running/accessible

-vb / Verify ALL Backup Servers
This should verify that ALL Backup License Servers are running/accessible

-vb SERVERNAME / Verify Specified Backup Server
This should verify that the SPECIFIED Backup Server is running.

The reason we need all of these switches is it will allow us to build canaries:

" Canary Tests are minimal tests to quickly and automatically verify that the everything you depend on is ready. You run Canary tests before other time-consuming tests, and before wasting time investigating in your code when the other tests are red. If the canary test fails, you know you have to fix something on the environments first. "

From https://dzone.com/articles/canary-tests

Against the licensing server to help operations teams understand troubles internally.

We need these to return non-zero exit codes so scripting around them is as simple as executing lmu -v and not wasting time parsing unparsable code (parsing for magic strings is the lowest of the low in devops world and indicates poor design of the tooling).

7 Comments | Posted by Ace Olszowka to Licensing on 10/26/2018 3:48 PM
Ace Olszowka
Come in today; find the backup license server has disconnected itself for some reason, no clue when.

Sure would have been nice to have some canary tests to know that this was broken and when...

** EDIT

"The primary server checks continuously to ensure that the backup server is present. If the primary cannot detect the backup for a period of 24 hours, it disconnects itself from the backup and propagates the change to all the clients."

(Bold and underlying mine). You get zero notification of this in the event logs, the only way you find it is if you look, and you gotta look manually because there is no canary. Good Luck to all those playing!

11/19/2018 3:01 PM   0  
Ace Olszowka
And again, we find ourselves in this situation; 98 days (or 3 Months, 6 days later) license server gone; backup license server long since disconnected; when did this happen? Who knows? Was there logging? "Nah that's never going to be used by anyone". A canary? That thing was a myth on the level of the loch ness monster.

When your business depends on licensing (or any other form of DRM) its probably a good idea to get it right; otherwise you get to deal with a bunch of unproductive workers and a lot of angry customers. You are also not really endearing developers who are looking to implement your technology. At the end of the day you gotta eat too, so its a necessary evil, but man you gotta get it right.

I want to really understand what other customer’s customers are doing: Is anyone else using the backup license server? How often are these systems being tested? Is no one else encountering this? Is no one putting processes in place to proactively defend against this?

L Peter Deutsch (of Sun Microsystems) identified all of these problems back in 1994; why must we continue to make the same mistakes of the past?

The fallacies are:
  1.     The network is reliable.
  2.     Latency is zero.
  3.     Bandwidth is infinite.
  4.     The network is secure.
  5.     Topology doesn't change.
  6.     There is one administrator.
  7.     Transport cost is zero.
  8.     The network is homogeneous.
https://en.wikipedia.org/wiki/Fallacies_of_distributed_computing

Pic unrelated:
User-added image

2/25/2019 2:55 PM   0  
Phillip Bratt
Wanted to mention that under tracker https://synergexresourcecenter.force.com/SiteTrackerDetail?id=38111 we are adding a canary and switching extending out the disconnect/lose of configuration setting to much longer (7 days). Just in case anyone is looking at this post. The rest of it is still under consideration but those parts of it we have plans to change.

4/10/2019 3:27 PM   0  
Ace Olszowka
Thanks Phil; Come to find out today the backup license server was again disconnected; who knows when; but sure was glad we didn't need it!

4/11/2019 10:24 PM   0  
Ace Olszowka
Dead again today; this time we made it 64 days; is there some type of timing bug?

Trying to channel my inter poet instead of my inner rage:
  
license server dead
performance of clients hurt
logs are of no use

Checks out with the Fish: http://writeahaiku.com/#license%20server%20dead%20%2F%20performance%20of%20clients%20hurt%20%2F%20logs%20are%20of%20no%20use

6/14/2019 3:10 PM   0  
Ace Olszowka
Dead again today; 24 days this time.

In talking with Support we do have several trackers open to try and get better logging.

7/8/2019 2:37 PM   0  
Ace Olszowka
We experienced a loss of our license server, because there is no canary (this ask), the backup license server de-auth'ed itself at some point in the past.

This resulted in downtime due to licensing failures. The root cause is still unknown, and the better logging was never implemented in REV11.

 

9/30/2020 9:26 PM   0  
Please log in to comment on this idea.