config/middleware.ini-dist has bad defaults

RESOLVED INVALID

Status

Socorro
General
RESOLVED INVALID
4 years ago
2 years ago

People

(Reporter: rhelmer, Assigned: lars)

Tracking

unspecified

Firefox Tracking Flags

(Not tracked)

Details

(Whiteboard: [qa-])

Attachments

(1 attachment)

(Reporter)

Description

4 years ago
There are at least a few problems:

* database_host (should be database_hostname)
* database_user (should be database_username)

Perhaps we should have a build step that generates a new dist file, and then fails if it doesn't match the currently checked-in one?
Lars is on it.
Assignee: nobody → lars
Status: NEW → ASSIGNED
(Assignee)

Comment 2

4 years ago
> Perhaps we should have a build step that generates a new dist file, and then fails if it
> doesn't match the currently checked-in one?

I've come to the conclusion that configman's automatic writing of config files is a useful starting place, but it is never good enough to be the final product.  Making an automated system of generating and testing would be impractical if the *.ini-dist included manual editing.
Component: Middleware → General
Target Milestone: --- → 66
(Assignee)

Comment 3

4 years ago
Created attachment 828727 [details] [review]
PR a re-imagining of all the default ini files.

I fully expect some back and forth and compromise on these files.  I made them as minimal as I could.
Attachment #828727 - Flags: review?(rhelmer)
(Assignee)

Comment 4

4 years ago
if these changes are accepted, then I intend on making another set of ini files for our prod/staging/dev environments based on these files.  Between those three environments, I hope to config most differences to the ./common_*.ini files.
(Assignee)

Comment 5

4 years ago
I find this very confusing.  We seem to have two opposing uses of the *.ini-dist configuration files.  

* They appear to be used in the Vagrant VM where HBase is not allowed.  
* They appear to be used in Jenkins where HBase is required.  

rhelmer: am I correct or just confused?
Flags: needinfo?(rhelmer)
(Reporter)

Comment 6

4 years ago
(In reply to K Lars Lohn [:lars] [:klohn] from comment #5)
> I find this very confusing.  We seem to have two opposing uses of the
> *.ini-dist configuration files.  
> 
> * They appear to be used in the Vagrant VM where HBase is not allowed.


Not vagrant specifically - the install doc says to copy middleware.ini-dist to middleware.ini and use that as a starting point, so it's expected to be non-hbase (since we don't document HBase as a requirement anymore)


> * They appear to be used in Jenkins where HBase is required.  


Actually Jenkins does not connect to HBase at all as far as I know - only dev/stage/prod actually have hbase available.


> rhelmer: am I correct or just confused?
Flags: needinfo?(rhelmer)
(Assignee)

Comment 7

4 years ago
I think the conclusion is that I'm confused.  Well, the PR has some very nice ini files for use in the Vagrant VM.  It doesn't solve the problem in this bug, but they're very nice anyway.  I'm withdrawing the PR for use on some other bug someday.
(Assignee)

Updated

4 years ago
Target Milestone: 66 → 67
(Reporter)

Updated

4 years ago
Attachment #828727 - Flags: review?(rhelmer)

Updated

4 years ago
Whiteboard: [qa-]
middleware config is done with consulate these days.
Status: ASSIGNED → RESOLVED
Last Resolved: 2 years ago
Resolution: --- → INVALID
You need to log in before you can comment on or make changes to this bug.