Stress Test coming up soon

Greetings all,

I have ‘the works’ package coupled to an internal MSSQL installation within our facility. We have a fairly large internal network that Is running a web based document server and a couple outsourced applications that are storing data to the database. I have been developing with Igniton for just a few months. The project I am developing is about to ‘go live’ with several management users. It has been a struggle to truly convince the IT department of the stability of the server and my project.

Well, coming up in a week is what one of the IT guys is calling a stress test. They want to have every user launch the project to ‘see’ how it is going to behave. I have seen a couple other forum posts about ‘sticking up’ for Ignition. several issues that I know I have had to stress through have been the ‘unknown’ tables that Ignition has created. - Thank you Travis & Support for the nice breakdown list for the tables.
Another stressful issue has been the dis-trust of java itself. Microsoft seems to be blaming allot of malware on Java lately. We have other corporate level software that runs on Java so I am somewhat safe on that one. I have been somewhat lucky there for I have not had to do anything for most of the clients to actually launch.

Unfortunately I am nervous about the test. I did the wrong thing and mentioned what event that I felt would be the ‘back breaker’ if there was going to be one. Here is where I am looking for insight/experience from you guys and IA. I have a fairly large project that has a graphic with almost 90 air handlers in which each one can launch a pop up where I have embedded the almost dreadful active-x browser that will pull up an internally hosted webpage from the hvac controller itself. This one worries me for multiple instances hitting that server.

Another issue would be a particular easy chart that I have developed from pulling modbus data from power monitors into Ignition, Ignition created these tables for the charts in which I have 7 compressors and about a month worth of data. Some of the early testers ‘liked’ the dynamic “apply” feature where they can just un-check the compressor that they do not want on the chart. One thing that I think might bog down would be the event of ‘every user’ manipulating their charts at the same time.

Overall these are the only stressful events that might cause any issue. These are ONLY my opinions. I am looking for info here. I also would like to be watching the ‘status’ info on the homepage but i am more concerned about what IT is monitoring. They will likely be monitoring network traffic and memory/cpu load on the server machine itself.
I think I at least have one of those covered because I did go ahead and use two of the tags that are available from the gateway for CPU utilization and memory usage They are on the bottom of the menu displayed as led indicators on the project navigation menu.

I am not sure what role they want me to have in this test since it is somewhat a defensive role that I seem to be having to take. I would like to know what to watch from a designer standpoint? Are some of my thoughts and opinions accurate/inaccurate??
All insight/information is very much appreciated.
I could post a gateway backup but as of now I think even some of the labels on my screens would violate our security policy. I could maybe send it to IA for them to evaluate…ALL TIPS/POINTERS appreciated…

My advice is to embrace this as an important milestone during your testing. Avoid presenting it as an “end all” determiner to project success or failure. If configuration issues are found they will be dealt with. This is different from a final acceptance/sign off test where everything is complete. IMO (once you get more forum feedback than mine), get back to IT with something like, “I spoke with the vendor, they think a stress test is a great idea. I appreciate your support. Here’s what I expect. These are the concerns I have. I do/don’t anticipate needing additional support.”

How many users do you have? Is this similar to how you intend on eventually running? When is the stress test planned for? When do you go live? What are the specs of your Ignition/DB server(s)? Have you tried opening lots of instances of the project on a few machines? What did your system performance metrics look like?

You might consider writing a short “how to” use the system with screenshots (quick Powerpoint presentation, Word document or pdf would suffice). On the HVAC page of the tutorial, put a little exclamation/tip icon saying “limited concurrent connections to HVAC controller” or something like that. I’d do you homework with the docs/vendor to see if you can get some info. You can also try to “stress” that yourself, seeing what the recovery process is and how it behaves if you can slam it. You might do something similar for the graphing page.

You should test it yourself beforehand, to the best of your ability. I’m thinking that your role will be monitoring the system on the Gateway for number of users versus resource utilization. Unless it’s a total flop, you can then report it to management as a success with metrics to back it up. This should even be the case if something specific doesn’t work - record that as a “lesson learned” - you’ll go back and address that.

Best of luck. Please keep us posted on how it goes.

good luck. Ignition is a great package and I dont believe you will have many issues. Just make sure and document everything and dont let an unrealistic approach be the deciding factor on if the testing was a success or not.

Man, now my brain’s on overdrive. I’m going to be the philosophical post in this thread. In my opinion, the spiritual (or mental, if you prefer) aspect of this is just as important. I’d even go so far to say that it’s even more so.

I agree with Nathan in that this is a milestone, not a stumbling block. And it certainly should not be a cause for worry.

All good expository writing has three points, so here they come. :slight_smile:

  1. Worry is unhelpful. It doesn’t change the past, and it can’t affect the future. But it can make you feel miserable.

  2. Worry is unrealistic. The more we worry over something the worse the problem seems to get. It’s where the “making a mountain out of a molehill” phrase comes from.

  3. Worry is unhealthy. Headaches, backaches, tightness in the chest, etc.-- our bodies just weren’t made for worrying.

So, Be sure to remind everyone involved (including yourself) that this is only a test. Since everyone is supposed to be on the same team, it’s not a win/lose, or even a pass/fail (unless, of course, it’s a spectacular Mythbusters style failure-- then you can spin it as an undocumented feature). Seriously, though, all this should do is find out where any weak spots so you can make a bigger better application. :smiley:

Greetings all,

Thank you for the advice!
Looks like the test is on Thursday. I will try not to be ‘worried’.
I had a nice talk with management about software vs. hardware. i used the example from the spc webinar
where the question was posed and the IA guys mentioned that there was one of their clients that has up to 200 clients.
I also mentioned that it will be more of an indication of our hardware/network architecture weaknesses than of the application. Though there are things that I can do to lessen the impacts. Which I am thankful for.
I may very well put the ‘apply button’ back in on the charts, and state some of the limitations of the HVAC controller ‘on screen’. I can also alter the tag ‘poll rates’ which is a feature that i am also thankful for.
I will post back results depending what role IT lets me take. Just wanted to do a quick post to say thank you!

I am posting this all from my Raspberry Pi… I still have not got Ignition running on it or a client launched.

Unlike the gooseberry I think the cotton candy may be able to run a client…

Hey, tailfire! How did the test go?

I was wondering the same.

I wish I had some other news than what I do.
The test did not happen, I am not sure if it will.
Should I go ahead and try to follow up promoting it, just for the experience and insight that can be gained?
I think there were ‘other political’ type influences that were driving this, but now i would just like the experience… :/

You might still consider pushing it on your own terms. It sounds like there was value to determine what kind of load the system could reasonably maintain. Are you getting the level of support from IT and management that you need?

In light of the fact that this is a somewhat public forum; I get
all the help and loving support that I need from my IT and Management… :open_mouth:

:mrgreen: Yeah, that’s the ticket

Great to hear! I’m still interested in your results as they come in.

[quote=“tailfire”]In light of the fact that this is a somewhat public forum; I get
all the help and loving support that I need from my IT and Management… :open_mouth:[/quote]