jberkus

Open Sourcing an Internal Project

VM

Author of fossforge.com (book in beta)

Releasing software is just the start.  If you want to get stuff, throwing code over the wall  won't get you it.  Guides you find for "how companies OSS" are too generic, you need to figure out what's specific to your company.  

Slides: archive.org/details/oscon2018-internalproject

jberkus

What	Who	APR	MAY	JUN	JUL	DEV WEEK	TEST GATES
Start of Release Cycle	Lead	2				week 1
Finalize Schedule	Lead	4
Begin collecting planned work from SIGs	Lead, Features Lead	9				week 2
Begin weekly release team meetings	Lead	10
Begin weekly status reports at Community	Lead, Shadow	12
Finalize Release Team	Lead	15
Start Release Notes Draft	Release Notes Lead	17				week 3

jberkus

Bug Wrassler Release Team Role

Your job for the release is to make sure that bugs (issues) and their related PRs which affect the release are dealt with in a timely fashion. In general, you will be polling bugs, checking status, sending reminders and questions to contributors and SIG leads, and publishing summary reports.

Secondarily, you will be helping improve automation around issue management and release tagging.

How this works depends on where you are in the release cycle. There are five relevant periods where your workload changes:

1. Early Release: between Feature Freeze and a week before Code Slush.

jberkus

-- check for exact matches
WITH index_cols_ord as (
    SELECT attrelid, attnum, attname
    FROM pg_attribute
        JOIN pg_index ON indexrelid = attrelid
    WHERE indkey[0] > 0
    ORDER BY attrelid, attnum
),
index_col_list AS (
    SELECT attrelid,

jberkus

-- check for containment
-- i.e. index A contains index B
-- and both share the same first column
-- but they are NOT identical

WITH index_cols_ord as (
    SELECT attrelid, attnum, attname
    FROM pg_attribute
        JOIN pg_index ON indexrelid = attrelid
    WHERE indkey[0] > 0

jberkus

SELECT * FROM (
SELECT
  schemaname, tablename, 
  ROUND(CASE WHEN otta=0 THEN 0.0 ELSE ((sml.relpages/otta::numeric) * 100 - 100) END) AS tbloat,
  CASE WHEN relpages < otta THEN 0 ELSE round(bs*(sml.relpages-otta)::numeric / (1024^2)::numeric , 2 ) END AS wastedmb,
  ROUND(AVG(CASE WHEN iotta >= ipages THEN 0 
  		 WHEN iotta = 0 THEN 0 
		 ELSE (ipages/iotta::numeric) * 100 - 100 END), -1) AS idxbloat,
  SUM(CASE WHEN ipages < iotta THEN 0 ELSE round( bs*(ipages-iotta)::numeric / (1024^2)::numeric, 2 ) END) AS wastedidxmb
FROM (

jberkus

WITH 
write_adjust AS (
    -- change the below to 1.0 if pg_stats goes back to
    -- the creation of the database
    SELECT 0.0 AS adjustment
),
index_usage AS (
    SELECT  sut.relid,
            current_database() AS database,
            sut.schemaname::text as schema_name, 

jberkus

#!/bin/bash

NUMMIN=$(($1 * 2 + 10))
CURMIN=0

while [ $CURMIN -lt $NUMMIN ] 
do

    psql -q -t -A -c "SELECT now() as ts, state, client_addr, count(*) as num_conn, avg(now() - query_start) as avg_query, \
max(now() - query_start) as max_query, avg(now() - xact_start) as avg_xact, \

jberkus

Steering Committee Update

Showed list of Steering Committee backlog.

We had a meeting, and the two big items we'd been pushing on hard were:

• votining in the proposals in the bootstrap committee
• how we're going to handle incubator and contrib etc.

jberkus

Demos in the booth
Virtualization Salon
OpenShift Commons
Talks?

Dinner/other team outing?