From a9046f3c45ced7237aaeacd61b413b8197359afd Mon Sep 17 00:00:00 2001
From: Peter Dillinger <peterd@fb.com>
Date: Tue, 2 Mar 2021 11:40:36 -0800
Subject: [PATCH] Revamp check_format_compatible.sh (#8012)

Summary:
* Adds backup/restore forward/backward compatibility testing
* Adds forward/backward compatibility testing to sst ingestion
* More structure sharing and comments for the lists of branches
comprising each group
* Less reliant on invariants between groups with de-duplication logic
* Restructured for n+1 branch checkout+build steps rather than something
like 3n. Should be much faster despite more checks.

And to make manual runs easier

* On success, restores working trees to original working branch (aborts
early if uncommitted changes) and deletes temporary branch & remote
* Adds SHORT_TEST=1 mode that uses only the oldest version for each
* Adds USE_SSH=1 to use ssh instead of https for github
group

Pull Request resolved: https://github.com/facebook/rocksdb/pull/8012

Test Plan:
a number of manual tests, mostly with SHORT_TEST=1. Using one
version older for any of the groups (except I didn't check
db_backward_only_refs) fails. Changing default format_version to 5
(planned) without updating this script fails as it should, and passes
with appropriate update. Full local run passed (had to remove "2.7.fb.branch"
due to compiler issues, also before this change).

Reviewed By: riversand963

Differential Revision: D26735840

Pulled By: pdillinger

fbshipit-source-id: 1320c22de5674760657e385aa42df9fade8b6fff
---
 tools/backup_db.sh               |  15 ++
 tools/check_format_compatible.sh | 334 +++++++++++++++++++++++--------
 tools/restore_db.sh              |  15 ++
 tools/write_external_sst.sh      |   1 +
 4 files changed, 282 insertions(+), 83 deletions(-)
 create mode 100755 tools/backup_db.sh
 create mode 100755 tools/restore_db.sh

diff --git a/tools/backup_db.sh b/tools/backup_db.sh
new file mode 100755
index 000000000..aa82f1dba
--- /dev/null
+++ b/tools/backup_db.sh
@@ -0,0 +1,15 @@
+#!/usr/bin/env bash
+# Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved.
+#
+#
+
+if [ "$#" -lt 2 ]; then
+  echo "usage: ${BASH_SOURCE[0]} <DB Path> <Backup Dir>"
+  exit 1
+fi
+
+db_dir="$1"
+backup_dir="$2"
+
+echo "== Backing up DB $db_dir to $backup_dir"
+./ldb backup --db="$db_dir" --backup_dir="$backup_dir"
diff --git a/tools/check_format_compatible.sh b/tools/check_format_compatible.sh
index c7d86027e..1d4076995 100755
--- a/tools/check_format_compatible.sh
+++ b/tools/check_format_compatible.sh
@@ -1,25 +1,83 @@
 #!/usr/bin/env bash
 # Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved.
 #
-# A shell script to load some pre generated data file to a DB using ldb tool
-# ./ldb needs to be avaible to be executed.
+# A shell script to build and run different versions of ldb to check for
+# expected forward and backward compatibility with "current" version. The
+# working copy must have no uncommitted changes.
+#
+# Usage: <SCRIPT> [ref_for_current]
+# `ref_for_current` can be a revision, tag, commit or branch name. Default is HEAD.
 #
-# Usage: <SCRIPT> [checkout]
-# `checkout` can be a tag, commit or branch name. Will build using it and check DBs generated by all previous branches (or tags for very old versions without branch) can be opened by it.
 # Return value 0 means all regression tests pass. 1 if not pass.
+#
+# Environment options:
+#  SHORT_TEST=1 - Test only the oldest branch for each kind of test. This is
+#    a good choice for PR validation as it is relatively fast and will find
+#    most issues.
+#  USE_SSH=1 - Connect to GitHub with ssh instead of https
 
-scriptpath=`dirname $BASH_SOURCE`
-test_dir=${TEST_TMPDIR:-"/tmp"}"/format_compatible_check"
+if ! git diff-index --quiet HEAD; then
+  echo "You have uncommitted changes. Aborting."
+  exit 1
+fi
+
+current_checkout_name=${1:-HEAD}
+# This allows the script to work even if with transient refs like "HEAD"
+current_checkout_hash="$(git rev-parse --quiet --verify $current_checkout_name)"
+
+if [ "$current_checkout_hash" == "" ]; then
+  echo "Not a recognized ref: $current_checkout_name"
+  exit 1
+fi
+
+# To restore to prior branch at the end
+orig_branch="$(git rev-parse --abbrev-ref HEAD)"
+tmp_branch=_tmp_format_compatible
+tmp_origin=_tmp_origin
+
+# Don't depend on what current "origin" might be
+set -e
+git remote remove $tmp_origin 2>/dev/null || true
+if [ "$USE_SSH" ]; then
+  git remote add $tmp_origin "git@github.com:facebook/rocksdb.git"
+else
+  git remote add $tmp_origin "https://github.com/facebook/rocksdb.git"
+fi
+git fetch $tmp_origin
+
+cleanup() {
+  echo "== Cleaning up"
+  git checkout "$orig_branch" || true
+  git branch -D $tmp_branch || true
+  git remote remove $tmp_origin || true
+}
+trap cleanup EXIT # Always clean up, even on failure or Ctrl+C
+
+scriptpath=`dirname ${BASH_SOURCE[0]}`
+
+test_dir=${TEST_TMPDIR:-"/tmp"}"/rocksdb_format_compatible_$USER"
+rm -rf ${test_dir:?}
+
+# For saving current version of scripts as we checkout different versions to test
 script_copy_dir=$test_dir"/script_copy"
-input_data_path=$test_dir"/test_data_input/"
+mkdir -p $script_copy_dir
+cp -f $scriptpath/*.sh $script_copy_dir
+
+# For shared raw input data
+input_data_path=$test_dir"/test_data_input"
+mkdir -p $input_data_path
+# For external sst ingestion test
+ext_test_dir=$test_dir"/ext"
+mkdir -p $ext_test_dir
+# For DB dump test
+db_test_dir=$test_dir"/db"
+mkdir -p $db_test_dir
+# For backup/restore test (uses DB test)
+bak_test_dir=$test_dir"/bak"
+mkdir -p $bak_test_dir
 
 python_bin=$(which python3 || which python || echo python3)
 
-mkdir $test_dir || true
-mkdir $input_data_path || true
-rm -rf $script_copy_dir
-cp $scriptpath $script_copy_dir -rf
-
 # Generate random files.
 for i in {1..6}
 do
@@ -55,11 +113,55 @@ with open('${sorted_input_data}', 'w') as f:
     print(k + " ==> " + v, file=f)
 EOF
 
-declare -a backward_compatible_checkout_objs=("2.2.fb.branch" "2.3.fb.branch" "2.4.fb.branch" "2.5.fb.branch" "2.6.fb.branch" "2.7.fb.branch" "2.8.1.fb" "3.0.fb.branch" "3.1.fb" "3.2.fb" "3.3.fb" "3.4.fb" "3.5.fb" "3.6.fb" "3.7.fb" "3.8.fb" "3.9.fb" "4.2.fb" "4.3.fb" "4.4.fb" "4.5.fb" "4.6.fb" "4.7.fb" "4.8.fb" "4.9.fb" "4.10.fb" "4.11.fb" "4.12.fb" "4.13.fb" "5.0.fb" "5.1.fb" "5.2.fb" "5.3.fb" "5.4.fb" "5.5.fb" "5.6.fb" "5.7.fb" "5.8.fb" "5.9.fb" "5.10.fb" "5.11.fb" "5.12.fb" "5.13.fb" "5.14.fb" "5.15.fb")
-declare -a forward_compatible_checkout_objs=() # N/A at the moment
-declare -a forward_compatible_with_options_checkout_objs=("5.16.fb" "5.17.fb" "5.18.fb" "6.0.fb" "6.1.fb" "6.2.fb" "6.3.fb" "6.4.fb" "6.5.fb" "6.6.fb" "6.7.fb" "6.8.fb" "6.9.fb" "6.10.fb" "6.11.fb" "6.12.fb" "6.13.fb" "6.14.fb" "6.15.fb" "6.16.fb" "6.17.fb")
-declare -a checkout_objs=(${backward_compatible_checkout_objs[@]} ${forward_compatible_checkout_objs[@]} ${forward_compatible_with_options_checkout_objs[@]})
-declare -a extern_sst_ingestion_compatible_checkout_objs=("5.16.fb" "5.17.fb" "5.18.fb" "6.0.fb" "6.1.fb" "6.2.fb" "6.3.fb" "6.4.fb" "6.5.fb" "6.6.fb" "6.7.fb" "6.8.fb" "6.9.fb" "6.10.fb" "6.11.fb" "6.12.fb" "6.13.fb" "6.14.fb" "6.15.fb" "6.16.fb" "6.17.fb")
+# db_backward_only_refs defined below the rest
+
+# To check for DB forward compatibility with loading options (old version
+# reading data from new), as well as backward compatibility
+declare -a db_forward_with_options_refs=("5.16.fb" "5.17.fb" "5.18.fb" "6.0.fb" "6.1.fb" "6.2.fb" "6.3.fb" "6.4.fb" "6.5.fb" "6.6.fb" "6.7.fb" "6.8.fb" "6.9.fb" "6.10.fb" "6.11.fb" "6.12.fb" "6.13.fb" "6.14.fb" "6.15.fb" "6.16.fb" "6.17.fb")
+# To check for DB forward compatibility without loading options (in addition
+# to the "with loading options" set), as well as backward compatibility
+declare -a db_forward_no_options_refs=() # N/A at the moment
+
+# To check for SST ingestion backward compatibility (new version reading
+# data from old) (ldb ingest_extern_sst added in 5.16.x, back-ported to
+# 5.14.x, 5.15.x)
+declare -a ext_backward_only_refs=("5.14.fb" "5.15.fb")
+# To check for SST ingestion forward compatibility (old version reading
+# data from new) as well as backward compatibility
+declare -a ext_forward_refs=("${db_forward_no_options_refs[@]}" "${db_forward_with_options_refs[@]}")
+
+# To check for backup backward compatibility (new version reading data
+# from old) (ldb backup/restore added in 4.11.x)
+declare -a bak_backward_only_refs=("4.11.fb" "4.12.fb" "4.13.fb" "5.0.fb" "5.1.fb" "5.2.fb" "5.3.fb" "5.4.fb" "5.5.fb" "5.6.fb" "5.7.fb" "5.8.fb" "5.9.fb" "5.10.fb" "5.11.fb" "5.12.fb" "5.13.fb" "${ext_backward_only_refs[@]}")
+# To check for backup forward compatibility (old version reading data
+# from new) as well as backward compatibility
+declare -a bak_forward_refs=("${db_forward_no_options_refs[@]}" "${db_forward_with_options_refs[@]}")
+
+# Branches (git refs) to check for DB backward compatibility (new version
+# reading data from old) (in addition to the "forward compatible" list)
+# NOTE: 2.7.fb.branch shows assertion violation in some configurations
+declare -a db_backward_only_refs=("2.2.fb.branch" "2.3.fb.branch" "2.4.fb.branch" "2.5.fb.branch" "2.6.fb.branch" "2.7.fb.branch" "2.8.1.fb" "3.0.fb.branch" "3.1.fb" "3.2.fb" "3.3.fb" "3.4.fb" "3.5.fb" "3.6.fb" "3.7.fb" "3.8.fb" "3.9.fb" "4.2.fb" "4.3.fb" "4.4.fb" "4.5.fb" "4.6.fb" "4.7.fb" "4.8.fb" "4.9.fb" "4.10.fb" "${bak_backward_only_refs[@]}")
+
+if [ "$SHORT_TEST" ]; then
+  # Use only the first (if exists) of each list
+  db_backward_only_refs=(${db_backward_only_refs[0]})
+  db_forward_no_options_refs=(${db_forward_no_options_refs[0]})
+  db_forward_with_options_refs=(${db_forward_with_options_refs[0]})
+  ext_backward_only_refs=(${ext_backward_only_refs[0]})
+  ext_forward_refs=(${ext_forward_refs[0]})
+  bak_backward_only_refs=(${bak_backward_only_refs[0]})
+  bak_forward_refs=(${bak_forward_refs[0]})
+fi
+
+# De-duplicate & accumulate
+declare -a checkout_refs=()
+for checkout_ref in "${db_backward_only_refs[@]}" "${db_forward_no_options_refs[@]}" "${db_forward_with_options_refs[@]}" "${ext_backward_only_refs[@]}" "${ext_forward_refs[@]}" "${bak_backward_only_refs[@]}" "${bak_forward_refs[@]}"
+do
+  if [ ! -e $db_test_dir/$checkout_ref ]; then
+    mkdir -p $db_test_dir/$checkout_ref
+    checkout_refs+=($checkout_ref)
+  fi
+done
 
 generate_db()
 {
@@ -105,89 +207,155 @@ ingest_external_sst()
     set -e
 }
 
-# Sandcastle sets us up with a remote that is just another directory on the same
-# machine and doesn't have our branches. Need to fetch them so checkout works.
-# Remote add may fail if added previously (we don't cleanup).
-git remote add github_origin "https://github.com/facebook/rocksdb.git"
-set -e
-git fetch github_origin
+backup_db()
+{
+    set +e
+    $script_copy_dir/backup_db.sh $1 $2
+    if [ $? -ne 0 ]; then
+        echo ==== Error backing up DB $1 to $2 ====
+        exit 1
+    fi
+    set -e
+}
 
-# Compatibility test for external SST file ingestion
-for checkout_obj in "${extern_sst_ingestion_compatible_checkout_objs[@]}"
-do
-  echo == Generating DB with extern SST file in "$checkout_obj" ...
-  git checkout github_origin/$checkout_obj -b $checkout_obj
-  make clean
-  DISABLE_WARNING_AS_ERROR=1 make ldb -j32
-  write_external_sst $input_data_path $test_dir/$checkout_obj $test_dir/$checkout_obj
-  ingest_external_sst $test_dir/$checkout_obj $test_dir/$checkout_obj
-done
+restore_db()
+{
+    set +e
+    $script_copy_dir/restore_db.sh $1 $2
+    if [ $? -ne 0 ]; then
+        echo ==== Error restoring from $1 to $2 ====
+        exit 1
+    fi
+    set -e
+}
 
-checkout_flag=${1:-"master"}
+member_of_array()
+{
+  local e match="$1"
+  shift
+  for e; do [[ "$e" == "$match" ]] && return 0; done
+  return 1
+}
 
-echo == Building $checkout_flag debug
-git checkout github_origin/$checkout_flag -b tmp-$checkout_flag
+# General structure from here:
+# * Check out, build, and do stuff with the "current" branch.
+# * For each older branch under consideration,
+#   * Check out, build, and do stuff with it, potentially using data
+#     generated from "current" branch.
+# * (Again) check out, build, and do (other) stuff with the "current"
+#    branch, potentially using data from older branches.
+#
+# This way, we only do at most n+1 checkout+build steps, without the
+# need to stash away executables.
+
+# Decorate name
+current_checkout_name="$current_checkout_name ($current_checkout_hash)"
+
+echo "== Building $current_checkout_name debug"
+git checkout -B $tmp_branch $current_checkout_hash
 make clean
 DISABLE_WARNING_AS_ERROR=1 make ldb -j32
-compare_base_db_dir=$test_dir"/base_db_dir"
-write_external_sst $input_data_path $compare_base_db_dir $compare_base_db_dir
-ingest_external_sst $compare_base_db_dir $compare_base_db_dir
 
-for checkout_obj in "${extern_sst_ingestion_compatible_checkout_objs[@]}"
+echo "== Using $current_checkout_name, generate DB with extern SST and ingest"
+current_ext_test_dir=$ext_test_dir"/current"
+write_external_sst $input_data_path ${current_ext_test_dir}_pointless $current_ext_test_dir
+ingest_external_sst ${current_ext_test_dir}_ingest $current_ext_test_dir
+
+echo "== Generating DB from $current_checkout_name ..."
+current_db_test_dir=$db_test_dir"/current"
+generate_db $input_data_path $current_db_test_dir
+
+echo "== Creating backup of DB from $current_checkout_name ..."
+current_bak_test_dir=$bak_test_dir"/current"
+backup_db $current_db_test_dir $current_bak_test_dir
+
+for checkout_ref in "${checkout_refs[@]}"
 do
-  echo == Build "$checkout_obj" and try to open DB generated using $checkout_flag
-  git checkout $checkout_obj
+  echo "== Building $checkout_ref debug"
+  git reset --hard $tmp_origin/$checkout_ref
   make clean
   DISABLE_WARNING_AS_ERROR=1 make ldb -j32
-  compare_db $test_dir/$checkout_obj $compare_base_db_dir db_dump.txt 1 1
-  git checkout tmp-$checkout_flag
-  # Clean up
-  git branch -D $checkout_obj
+
+  # We currently assume DB backward compatibility for every branch listed
+  echo "== Use $checkout_ref to generate a DB ..."
+  generate_db $input_data_path $db_test_dir/$checkout_ref
+
+  if member_of_array "$checkout_ref" "${ext_backward_only_refs[@]}" ||
+    member_of_array "$checkout_ref" "${ext_forward_refs[@]}"
+  then
+    echo "== Use $checkout_ref to generate DB with extern SST file"
+    write_external_sst $input_data_path $ext_test_dir/${checkout_ref}_pointless $ext_test_dir/$checkout_ref
+  fi
+
+  if member_of_array "$checkout_ref" "${ext_forward_refs[@]}"
+  then
+    echo "== Use $checkout_ref to ingest extern SST file and compare vs. $current_checkout_name"
+    ingest_external_sst $ext_test_dir/${checkout_ref}_ingest $ext_test_dir/$checkout_ref
+    compare_db $ext_test_dir/${checkout_ref}_ingest ${current_ext_test_dir}_ingest db_dump.txt 1 1
+
+    rm -rf ${ext_test_dir:?}/${checkout_ref}_ingest
+    echo "== Use $checkout_ref to ingest extern SST file from $current_checkout_name"
+    ingest_external_sst $ext_test_dir/${checkout_ref}_ingest $current_ext_test_dir
+    compare_db $ext_test_dir/${checkout_ref}_ingest ${current_ext_test_dir}_ingest db_dump.txt 1 1
+  fi
+
+  if member_of_array "$checkout_ref" "${db_forward_no_options_refs[@]}" ||
+    member_of_array "$checkout_ref" "${db_forward_with_options_refs[@]}"
+  then
+    echo "== Use $checkout_ref to open DB generated using $current_checkout_name..."
+    compare_db $db_test_dir/$checkout_ref $current_db_test_dir forward_${checkout_ref}_dump.txt 0
+  fi
+
+  if member_of_array "$checkout_ref" "${db_forward_with_options_refs[@]}"
+  then
+    echo "== Use $checkout_ref to open DB generated using $current_checkout_name with its options..."
+    compare_db $db_test_dir/$checkout_ref $current_db_test_dir forward_${checkout_ref}_dump.txt 1 1
+  fi
+
+  if member_of_array "$checkout_ref" "${bak_backward_only_refs[@]}" ||
+    member_of_array "$checkout_ref" "${bak_forward_refs[@]}"
+  then
+    echo "== Use $checkout_ref to backup DB"
+    backup_db $db_test_dir/$checkout_ref $bak_test_dir/$checkout_ref
+  fi
+
+  if member_of_array "$checkout_ref" "${bak_forward_refs[@]}"
+  then
+    echo "== Use $checkout_ref to restore DB from $current_checkout_name"
+    rm -rf ${db_test_dir:?}/$checkout_ref
+    restore_db $current_bak_test_dir $db_test_dir/$checkout_ref
+    compare_db $db_test_dir/$checkout_ref $current_db_test_dir forward_${checkout_ref}_dump.txt 0
+  fi
 done
 
-echo == Finish compatibility test for SST ingestion.
-
-for checkout_obj in "${checkout_objs[@]}"
-do
-   echo == Generating DB from "$checkout_obj" ...
-   git checkout github_origin/$checkout_obj -b $checkout_obj
-   make clean
-   DISABLE_WARNING_AS_ERROR=1 make ldb -j32
-   generate_db $input_data_path $test_dir/$checkout_obj
-done
-
-checkout_flag=${1:-"master"}
-
-echo == Building $checkout_flag debug
-git checkout tmp-$checkout_flag
+echo "== Building $current_checkout_name debug (again, final)"
+git reset --hard $current_checkout_hash
 make clean
 DISABLE_WARNING_AS_ERROR=1 make ldb -j32
-compare_base_db_dir=$test_dir"/base_db_dir"
-echo == Generate compare base DB to $compare_base_db_dir
-generate_db $input_data_path $compare_base_db_dir
 
-for checkout_obj in "${checkout_objs[@]}"
+for checkout_ref in "${checkout_refs[@]}"
 do
-   echo == Opening DB from "$checkout_obj" using debug build of $checkout_flag ...
-   compare_db $test_dir/$checkout_obj $compare_base_db_dir db_dump.txt 1 0
-done
+  # We currently assume DB backward compatibility for every branch listed
+  echo "== Use $current_checkout_name to open DB generated using $checkout_ref..."
+  compare_db $db_test_dir/$checkout_ref $current_db_test_dir db_dump.txt 1 0
 
-for checkout_obj in "${forward_compatible_checkout_objs[@]}"
-do
-   echo == Build "$checkout_obj" and try to open DB generated using $checkout_flag...
-   git checkout $checkout_obj
-   make clean
-   DISABLE_WARNING_AS_ERROR=1 make ldb -j32
-   compare_db $test_dir/$checkout_obj $compare_base_db_dir forward_${checkout_obj}_dump.txt 0
-done
+  if member_of_array "$checkout_ref" "${ext_backward_only_refs[@]}" ||
+    member_of_array "$checkout_ref" "${ext_forward_refs[@]}"
+  then
+    rm -rf ${ext_test_dir:?}/${checkout_ref}_ingest
+    echo "== Use $current_checkout_name to ingest extern SST file from $checkout_ref"
+    ingest_external_sst $ext_test_dir/${checkout_ref}_ingest $current_ext_test_dir
+    compare_db $ext_test_dir/${checkout_ref}_ingest ${current_ext_test_dir}_ingest db_dump.txt 1 1
+  fi
 
-for checkout_obj in "${forward_compatible_with_options_checkout_objs[@]}"
-do
-   echo == Build "$checkout_obj" and try to open DB generated using $checkout_flag with its options...
-   git checkout $checkout_obj
-   make clean
-   DISABLE_WARNING_AS_ERROR=1 make ldb -j32
-   compare_db $test_dir/$checkout_obj $compare_base_db_dir forward_${checkout_obj}_dump.txt 1 1
+  if member_of_array "$checkout_ref" "${bak_backward_only_refs[@]}" ||
+    member_of_array "$checkout_ref" "${bak_forward_refs[@]}"
+  then
+    echo "== Use $current_checkout_name to restore DB from $checkout_ref"
+    rm -rf ${db_test_dir:?}/$checkout_ref
+    restore_db $bak_test_dir/$checkout_ref $db_test_dir/$checkout_ref
+    compare_db $db_test_dir/$checkout_ref $current_db_test_dir db_dump.txt 1 0
+  fi
 done
 
 echo ==== Compatibility Test PASSED ====
diff --git a/tools/restore_db.sh b/tools/restore_db.sh
new file mode 100755
index 000000000..ed89794b2
--- /dev/null
+++ b/tools/restore_db.sh
@@ -0,0 +1,15 @@
+#!/usr/bin/env bash
+# Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved.
+#
+#
+
+if [ "$#" -lt 2 ]; then
+  echo "usage: ${BASH_SOURCE[0]} <Backup Dir> <DB Path>"
+  exit 1
+fi
+
+backup_dir="$1"
+db_dir="$2"
+
+echo "== Restoring latest from $backup_dir to $db_dir"
+./ldb restore --db="$db_dir" --backup_dir="$backup_dir"
diff --git a/tools/write_external_sst.sh b/tools/write_external_sst.sh
index 3b02d082f..be01ae022 100755
--- a/tools/write_external_sst.sh
+++ b/tools/write_external_sst.sh
@@ -12,6 +12,7 @@ input_data_dir=$1
 db_dir=$2
 extern_sst_dir=$3
 rm -rf $db_dir
+mkdir -p $extern_sst_dir
 
 set -e