Question

What has been your biggest FME Oops moment?

  • 4 December 2023
  • 15 replies
  • 13 views
inge
Rank
Badge +1

We (Red Geographics) often organize masterclasses to show and tell people on how to do something in FME.

Our last masterclass of the year is going to be the opposite.

In the Anti-masterclass we're going to tell people how NOT to FME.

 

It's gonna be a fun way to end the year and share some funny situations, oops-moments and things to avoid.

I'd love to get some examples from FME experts from all over the world.

What has been a blunder or a funny mistake you want to share that we can incorporate into our (Dutch) webinar?

15 replies

nielsgerrits
Rank
Badge +39

I have far too many of those. A few that come to mind.

  • Connect an Emailer to an outputport ejecting more then a few features, creating a mailbomb.
  • Use a SystemCaller to delete a temp folder, have unexpected input, attempt to clear root.
  • Skipping OTAP and ruining a production database.

redgeographics
Rank
Userlevel 5
Badge +32

I have far too many of those. A few that come to mind.

  • Connect an Emailer to an outputport ejecting more then a few features, creating a mailbomb.
  • Use a SystemCaller to delete a temp folder, have unexpected input, attempt to clear root.
  • Skipping OTAP and ruining a production database.

I haven't done the mailbomb myself but I did do a Denial-of-service attack with an HTTPCaller once 😅

nielsgerrits
Rank
Badge +39

I haven't done the mailbomb myself but I did do a Denial-of-service attack with an HTTPCaller once 😅

To be honest, this was a teammate, but that is close enough. The worst consequence was that we reached our message quota for 24 hours, resulting in users not receiving FME Server generated email...

 

And I have pushed the limits using HTTPCallers as well, not sure if I really brought something down that way.

inge
Rank
Badge +1

I haven't done the mailbomb myself but I did do a Denial-of-service attack with an HTTPCaller once 😅

Thanks for sharing! They're pretty epic tbh

hkingsbury
Rank
Userlevel 6
Badge +36

Not updating the writer to the DEV environment and instead dropping all features in the production table 😬

 

Also got our IP blocked for a few hours from a local agencies REST endpoint by making too many calls to it

panda
Rank
Userlevel 1
Badge +14

1) Processing about 5 milion records with cache on and almost crashed the IT server because space on disk is at critical limit. Got calls from IT to stop that workspace lol. Always Make sure your cache was on temporary disk that have tons of space available, or always turn the cache off after testing if not utilizing FME Flow!

 

2) Forgot to switch production connection to dev/testing connection and wrote a whole bunch of bogus data to the production environment.

 

nielsgerrits
Rank
Badge +39

Not updating the writer to the DEV environment and instead dropping all features in the production table 😬

 

Also got our IP blocked for a few hours from a local agencies REST endpoint by making too many calls to it

Ouch, I know that feeling when you realize you need to get busy with backups.

mark2atsafe
Rank
Userlevel 5
Badge +35

I have far too many of those. A few that come to mind.

  • Connect an Emailer to an outputport ejecting more then a few features, creating a mailbomb.
  • Use a SystemCaller to delete a temp folder, have unexpected input, attempt to clear root.
  • Skipping OTAP and ruining a production database.

I've done exactly the same thing with an Emailer. Thankfully I had only 20-30 features, so it wasn't a complete bomb!

mark2atsafe
Rank
Userlevel 5
Badge +35

Oh yes, lots of things.

  • My common mistake recently is to call an API using different parameters, but get the same output as before! I can spend an hour hitting the API again and again before I remember that the JSONFragmenter is pointing to the same local file that I used to set it up, and not the API response.
  • The most embarrassing mistake is to do something like that and then file a bug report with our developers because I think the transformer is broken. If I'm lucky I find my oops before they look into it. If I'm unlucky, I waste their time and they then very tactfully tell me that I'm an idiot. We don't mind new users having this happen. But when your job title is "FME Evangelist" you're expected to know everything!
  • Not saving a workspace when I first create it. Then, inevitably, FME crashes and my work is completely lost with not even an automatic backup file to restore. Save your workspace as soon as you create it!
  • I've done the Emailer bomb too. Especially when I put in a Sampler and think I've set it to "First 1 Feature" but instead it is "Every 1 Feature"!

 

warrendev
Rank
Badge +21

1) Processing about 5 milion records with cache on and almost crashed the IT server because space on disk is at critical limit. Got calls from IT to stop that workspace lol. Always Make sure your cache was on temporary disk that have tons of space available, or always turn the cache off after testing if not utilizing FME Flow!

 

2) Forgot to switch production connection to dev/testing connection and wrote a whole bunch of bogus data to the production environment.

 

😑

Whoops!

ctredinnick
Rank
Userlevel 4
Badge +17

A variant on the emailer bomb. I created a custom transformer to upload documents by API to council's document management system, and shared it internally without training. Other people then picked it up and didn't realise that it uploaded documents once per feature passed into it. There was a process being run of 'read asbuilt data, check and upload file' which at the end was uploading a separate copy of the file for each feature it contained.

joepk
Rank
Userlevel 4
Badge +17

Oh yes, lots of things.

  • My common mistake recently is to call an API using different parameters, but get the same output as before! I can spend an hour hitting the API again and again before I remember that the JSONFragmenter is pointing to the same local file that I used to set it up, and not the API response.
  • The most embarrassing mistake is to do something like that and then file a bug report with our developers because I think the transformer is broken. If I'm lucky I find my oops before they look into it. If I'm unlucky, I waste their time and they then very tactfully tell me that I'm an idiot. We don't mind new users having this happen. But when your job title is "FME Evangelist" you're expected to know everything!
  • Not saving a workspace when I first create it. Then, inevitably, FME crashes and my work is completely lost with not even an automatic backup file to restore. Save your workspace as soon as you create it!
  • I've done the Emailer bomb too. Especially when I put in a Sampler and think I've set it to "First 1 Feature" but instead it is "Every 1 Feature"!

 

This post made me save my workspace.

nielsgerrits
Rank
Badge +39

A variant on the emailer bomb. I created a custom transformer to upload documents by API to council's document management system, and shared it internally without training. Other people then picked it up and didn't realise that it uploaded documents once per feature passed into it. There was a process being run of 'read asbuilt data, check and upload file' which at the end was uploading a separate copy of the file for each feature it contained.

Classic :) It does what I told it to do, but this is not what I wanted it to do.

inge
Rank
Badge +1

Loving all there stories!

Foutmelding FMEThis is my own personal favorite

FME was really pleading to stop what I was doing.

Just can't remember what I was trying here..

ebygomm
Rank
Userlevel 2
Badge +21

Leaving inspectors active in a child workspace, then running that workspace a few thousand times with a workspace runner.

Reply


Community Stats

29,355
Posts
111,609
Replies
37,016
Members
Terms & ConditionsCookie settings
A product of
Safe Software respectfully acknowledges that we live, learn and work on the traditional and unceded territories of the Kwantlen, Katzie, and Semiahmoo First Nations.