Y502K - Failure 1.35.4.1
Today the WSLCB and MJ FailWhale/LeakData have once again demonstrated their incompetence and negligence.
Mary Mueller, the CIO of the LCB sent this out today (we received at 18:31 Pacific Time)
Good evening, gang, I know that we were hoping to get this out earlier, but I wanted to let you know that we did find a couple of defects in release 1.35.4.1. Due to the severity of the issue to be fixed, I have reluctantly approved this to roll to Test ASAP and Prod tonight. We found that nine fields are now not returning values for one of the inventory GET calls. Business has reviewed and believes that the defects are sufficiently low impact that we should send it out.
Part of the release will be detailed information about the defects and the workarounds. Again, in the future for Sev-1s or non-Sev 1s, I won’t approve the release with any defects, but I understand the urgency to get this out.
More info to come tonight from the tech team with the release notes.
Mary Gould Mueller
Chief Information Officer, Washington State Liquor and Cannabis Board
Notice the part where there are now nine fields in this data-system that have become broken by this "fix" they tried to deploy. Clearly, this release shouldn't have been allow to proceed however, Mary places responsiblity on the "Business" division -- whatever that is.
For reference here are the fields that were somehow lost in this update.
- created_at
- updated_at
- lab_retest_id
- deleted_at
- global_created_by_mme_id
- lab_results_file_path
- lab_results_attested -- this field is critical to work around their existing failure to handle QA properly.
- lab_results_date
- global_mme_id -- this field tells us who actually owns the inventory -- yes, LeafData lost who owns the inventory.
Then, at 20:28 Mary sends another message:
I have stopped the production deployment. The team is working on the fix to the missing fields right now and we’ll be deploying to Test and Stage at the same time late tonight. I’ll email you in the morning when we have the fix version tested and verified as passed.
It's good they stopped the release of the horrible broken crap, yet it still serves to demonstrate how underwhelming their performance has been over the last six months.
TL;DR
- Fix 1.35.4.1 broke nine data fields (but claimed to fix some other severe issue).
- The "Business" division decided this was acceptable, Mary had no responsibility.
- Two hours later Mary steps in to stop the release of this turd.
As a software engineer with 20+ years of experience this is a failure at the CIO level. The new breakage is what is called a regression, if you have a regression bug it means your new code is a pile of shit, see any rant by Linux on "don't break userspace" Mary should never have sent this through to the phantom "Business" division.
Also
If you're interested in cannabis software developement, we've got a nice little community going in this group https://groups.google.com/forum/#!forum/cannabis-api -- come, join us.