36

I'm posting this on Meta as I'm not sure it belongs in the normal SFSE.

After getting an Internal Salesforce Error in a complex bit of Apex V35 logic, I filed a Premier Support case as the underlying cause of the error was not ascertainable from the debug logs - hoping that SFDC could just look in their logs and tell me what the error was so I would have a clue (error was occurring between end of a before delete trigger and the non-execution of the after-delete trigger)

As you probably know, Tier 2 SFDC Developer Support is a filter to make sure you have a reproducible case before kicking it to Tier 3 if the case is not something already known about. All I wanted was to get the underlying error message which didn't seem like that would be too hard.

The first live call with Tier 2 simply acknowledged the case and said they'd get back to me. Having worked with Tier 2 before on other issues, I wasn't too optimistic of a speedy resolution. Meanwhile, I spent hours engaged in selective reduction of the code and various test scenarios until I finally came up with 8 lines of Execute Anonymous that would reliably reproduce the Internal Salesforce Error and, eventually, I deduced the root cause on my own - all of which I added to the Support Case comments.

When Tier 2 called back the next day, they acknowledged my case comments and basically said, 'if others find this issue, it will get to development and be fixed in some future release'

OK, I get that SFDC is a big complex set of software and there are probably a zillion pending items to work on that include bug fixes. I don't actually expect SFDC to fix this issue as it only applies to deletion of parent objects with ContentVersion children where the lookup relationship says don't allow deletes of parent object.

But, I would think that Premier Support customers who report an Internal Salesforce Error with a reproducible test case and the underlying cause should at least get a "Thank You for helping make Salesforce a better system" - perhaps from Developer Relations.

End of 'not feeling loved' - your experiences/thoughts?

3
  • 1
    And you have Premier support and everything! As just another customer, I say thank you! Did you manager to get a KI link out of the process? I'll more than happily pass it among colleagues to mark as "this affects me".
    – Christian Carter
    Jan 29, 2016 at 4:26
  • @christian. I'm still working on the KI.
    – cropredy
    Jan 29, 2016 at 5:43
  • Please take a look at my question and idea here: meta.salesforce.stackexchange.com/questions/2382/…
    – Patlatus
    Oct 26, 2016 at 14:36

6 Answers 6

Reset to default
34

Thank you. Seriously. As you probably know, I'm a developer evangelist, part of developer relations. Right now there's no mechanism for us (developer relations) to know when a developer in our community submits a gack report with a reproducible test case, but it's something that should definitely be acknowledged. I'm guessing what happened in this case is that tier 2 followed their process in closing the case, since it appears to be a rare edge case, and then moved straight on to the next case in the queue. (Too many cases in that sentence, but you get the drift!)

Maybe we should have a process for support to flag a case for some love from developer relations. Not sure what form that love should take... An exclusive 'Bugfinder General' t-shirt? Let me bounce this around the team a bit...

And, @crop1645, I notice you're in the Bay Area. Beer is on me if/when we run into each other!

7
  • 2
    metadaddy - Thank you for this nice answer. I feel loved now :-). Now that I know these incidents are called gacks, I found the blog post by Josh Kaplan at developer.salesforce.com/blogs/engineering/2015/02/gack.html and will follow his advice. I wish you well in making change happen. And ... I actually (briefly) introduced myself to you at DF 14 when you were running off the stage after a keynote from the lesser-but-still-big cheeses. Perhaps we'll meet again at a future DF.
    – cropredy
    Jan 28, 2016 at 16:49
  • 6
    I've a list of gacks that I keep around, but I myself hardly bother reporting them anymore, because it doesn't seem to do any good. Granted, there are some gacks that I would expect to be gacks, but I'm still usually surprised when code just goes off the deep end instead of saying something reasonable. For example, passing classes in execute anonymous to batch classes causes gacks. I understand why, but it'd be nice to have a simple "I'm sorry. I can't let you do that, Dave (or, Brian, in this case)." Or not being able to chain queueables that callout.
    – sfdcfox
    Jan 29, 2016 at 22:44
  • 2
    I've had this same and similar issues. I once found an exploitable security problem. Support asked me to post the issue to the ideaexchange so that it could be prioritized through the usual channels and I was flabbergasted but didn't really have any other recourse until I emailed it to a PM who I met at Dreamforce, which felt like breaking the rules...
    – Benj
    Feb 1, 2016 at 18:49
  • 6
    @Benj There is a super secret hard to find email address at the end of Vulnerability Reporting Policy that you can use rather than posting security problems to the ideaexchange or developer forumn.
    – Daniel Ballinger
    Feb 8, 2016 at 21:16
  • @DanielBallinger good to know - thanks!
    – Benj
    Feb 9, 2016 at 21:40
  • @Benj That's just CRAZY to post fixing a security exploitable problem to the idea exchange! Duh, let's advertise the problem exists so someone can exploit it in any org they want to! Its as though these people just don't stop and think about what they're instructing you to do. Either that, or it's symptomatic of just how tightly their hands are being tied in addressing these matters. I strongly suspect it's more a case of the latter; that they're simply not being empowered to handle them.
    – crmprogdev
    Feb 12, 2016 at 14:23
  • @metadaddy one of the tools I used sent me a t-shirt after submitting multiple bugs with clear steps to reproduce. it was a big boost to know they valued the time i spent. unfortunately it's the exact opposite with sfdc, the general flow is 1) deny it's a bug, 2) tell me i need premier support for dev stuff 3) repeatedly request a goto meeting for me to point out steps to reproduce are in the ticket 4) go silent for a week 5) finally get it up to r&d 6) raise a family of four while i'm waiting 7) r&d confirms bug 8) i beg to make it a known issue so people can acknowledge it :(
    – Ralph Callaway
    Jun 6, 2018 at 2:45
28

I've been there. I gave up trying to explain to someone in support that I had a reproducible internal server error, they kept insisting that my code was at fault.

I also know someone who found a major security issue with the system, reported it, and got offered a hoodie.

Generally I've never really had great interactions with support, it's been about 50/50 as to whether I get some useful help. That's why I prefer using this site, because people are more knoweledgeable about issues than a lot of internal staff are!

4
  • 11
    a hoodie for a major security flaw, sheesh.
    – cropredy
    Jan 27, 2016 at 22:58
  • 6
    One of the members of my team, who encountered a gack while we were building a managed package, was instructed by T2 support to contact the developer of the managed package... thanks a lot, Support.
    – Mark Pond
    Jan 28, 2016 at 19:14
  • 4
    @MarkPond Yeah I've had support tell me that I need to talk to me before.
    – Matt Lacey Mod
    Jan 28, 2016 at 22:23
  • 1
    This and the many recent threads for any other company would cause them to take a hard look at their processes and fix them. The overwhelming consensus is that SF support is less than ideal and we just deal with the bugs, they will eventually find them on their own (byproduct of the GACK system??). The comments about directing us to talk to ourself is priceless!!
    – Eric
    Nov 22, 2017 at 0:45
15

I know the feeling for sure. In general the lack of receptiveness to bug reporting is very frustrating. I've been told several times to use the Idea Exchange, which really doesn't make sense, as enhancements and bugs should be handled differently.

11

I think this is unfortunate. It's been my observation over the last several years as a developer that it's been increasingly difficult to get a case opened with Salesforce Support that doesn't get immediately closed, regardless of how valid it is.

Many of us work in a variety of different Orgs for short periods of time as contractors or subcontractors. Some, only work in DE Orgs developing Apps. We're generally not considered "customers" in the traditional sense by Salesforce. So when we encounter issues, we're frequently not able to open and file cases like a traditional customer can. I've at times filed them on behalf of a customer and been successful in doing that. At other times, I've had them immediately closed, even though they were legitimate issues with the platform that needed their attention (and in a number of cases still do).

I think that's unfortunate because we use the knowledge we have to assist a great number of Salesforce's customers. In my view, taking that stance is very short-sighted on their part. I don't expect to receive the same priority as a premiere support customer, but to be told that because I don't have premiere support, I can't even open a case? To me, that's not being at all helpful to the many developers who support LOTS of customers.

And if you're wondering, yes, I took a really deep breath so I could dial back my comments to keep them G-rated.

9

I had a similar scenario last year where I raised a case with steps to reproduce and after few rounds of discussion (which went on for days of course), the support team replied a very similar statement as you posted that if multiple users report the same issue, it will get prioritised and included in their future rollouts..

but after further discussions all i managed was to get it added to Known issues this at least helps someone who faces a similar one can report on the page and the issue becomes publicly trackable..

1
  • 4
    Getting something onto Known Issues is a challenge in and of itself requiring multiple exhortations and frequent uses of the expression "[this bug] doesn't pass the straight face test". See salesforce.stackexchange.com/questions/100627/… for an example where I had to use the above to make it a Known Issue
    – cropredy
    Jan 28, 2016 at 17:58
5

I have to say I have had a slightly different experience recently...no thank you or recognition as such but the two Spring16 bugs we found both were accepted. One has already been patched and the other likely before Spring16 goes live. I can't find a known issue for either so I might inquire as to why there isn't.

It hasn't been easy and took an awful lot of pressure/pushing to do (to the point it has decimated our change cycle). It may just be that my client is a large Salesforce customer combined with shouting very loudly.

The only other time I got a bug recognised (which again took a lot of cajoling) and got a number it kind of disappeared...

1
  • 1
    cool and kudos -- given that I had a workaround, I opted against going the 'elevated voice' route but if it were a bug that was throwing off our ability to deploy, I would have dug deep into my 'persuasive skills' as we too are a large SFDC customer
    – cropredy
    Jan 30, 2016 at 17:03

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .