Aurora and Nightly use new table, apply to search

VERIFIED FIXED in 2.3.2

Status

Socorro
General
VERIFIED FIXED
7 years ago
7 years ago

People

(Reporter: adrian, Assigned: adrian)

Tracking

({regression})

Trunk
2.3.2
regression

Firefox Tracking Flags

(Not tracked)

Details

(Assignee)

Description

7 years ago
Aurora and Nightly are now using the "new" table for version info. I'll apply that change to search with PostgreSQL and ElasticSearch. 

Will use that occasion to refactor PG to use the version_info service and fix bug 683370.

Updated

7 years ago
Assignee: nobody → adrian
Blocks: 699451
Keywords: regression
Target Milestone: --- → 2.3.2

Comment 1

7 years ago
Commit pushed to https://github.com/mozilla/socorro

https://github.com/mozilla/socorro/commit/f264829d2a6c497d77c41ec2b7148925fd414e43
Merge pull request #133 from AdrianGaudebert/699494-search-aurora-nightly

Fixes bug 699494 - Added support for Aurora and Nightly crashes to search

Updated

7 years ago
Status: NEW → RESOLVED
Last Resolved: 7 years ago
Resolution: --- → FIXED

Comment 2

7 years ago
Commit pushed to https://github.com/mozilla/socorro

https://github.com/mozilla/socorro/commit/231bb16dba202f6048d224f74f932f92e48a1cba
Fixes bug 699494 - Added support for Aurora and Nightly crashes to search with PostgreSQL.
Adrian, can you please comment on steps to verify this bug?

Offhand I would say that we should ensure that all release types ("release", "beta", "nightly", "aurora") work and return the correct number of results; how should we determine what the "correct number" is?

Maybe Josh could dig up some numbers from the reports table to cross-check?
(Assignee)

Comment 4

7 years ago
To fully test something like that (and everything that is highly data-related), we will need a test battery that knows the initial dataset. This is something we discussed with Matt Brandt and that we intend to discuss more with the team soon. 

For example, we could have a db with a very small set of crashes (say between 100 and 1,000 total), and know everything about those crashes: how many "firefox", how many "beta", how many "mac"... Then we could run a bunch of queries against the middleware and check the returned data correspond to the numbers and signatures we want. If it's not the case we will be sure that something is wrong. 

But, we're not there yet, so for the moment, the best way to test that is to query for anything that is Aurora and/or Nightly and see that there are data! :)
QA verified on stage. Results for (In reply to Adrian Gaudebert [:adrian] from comment #4)
>for the moment, the best way to test that is to query for anything that is Aurora
>and/or Nightly and see that there are data! :)

QA Verified on stage... there be data for for Aurora

> To fully test something like that (and everything that is highly
> data-related), we will need a test battery that knows the initial dataset.
> This is something we discussed with Matt Brandt and that we intend to
> discuss more with the team soon. 

I'm all for this, perhaps we can have a QA workweek as a team (Socorro peoples) and knock this out this quarter?
Status: RESOLVED → VERIFIED
Component: Socorro → General
Product: Webtools → Socorro
You need to log in before you can comment on or make changes to this bug.