From fbe68d516c9bf9ccf880d6be54883899a8b471d2 Mon Sep 17 00:00:00 2001
From: Jack Robison <jackrobison@lbry.io>
Date: Tue, 14 Jun 2022 13:55:53 -0400
Subject: [PATCH] optionally initialize rocksdb from a snapshot

fixes https://github.com/lbryio/hub/issues/10
---
 docs/docker_examples/docker-compose.yml |  3 ++-
 docs/docker_examples/hub-compose.yml    |  1 +
 scripts/entrypoint.sh                   | 16 ++++++++++++++++
 3 files changed, 19 insertions(+), 1 deletion(-)

diff --git a/docs/docker_examples/docker-compose.yml b/docs/docker_examples/docker-compose.yml
index cc979af..5d33924 100644
--- a/docs/docker_examples/docker-compose.yml
+++ b/docs/docker_examples/docker-compose.yml
@@ -17,6 +17,7 @@ services:
       - "lbry_rocksdb:/database"
     environment:
       - HUB_COMMAND=scribe
+      - SNAPSHOT_URL=https://snapshots.lbry.com/hub/lbry-rocksdb.zip
     command:  # for full options, see `scribe --help`
       - "--daemon_url=http://lbry:lbry@127.0.0.1:9245"
       - "--max_query_workers=2"
@@ -41,7 +42,7 @@ services:
       - "--elastic_port=9200"                 # elasticsearch port
       - "--elastic_notifier_host=127.0.0.1"   # address for the elastic sync notifier to connect to
       - "--elastic_notifier_port=19080"
-  scribe_hub:
+  herald:
     depends_on:
       - lbcd
       - scribe_elastic_sync
diff --git a/docs/docker_examples/hub-compose.yml b/docs/docker_examples/hub-compose.yml
index 2517e63..96e70f5 100644
--- a/docs/docker_examples/hub-compose.yml
+++ b/docs/docker_examples/hub-compose.yml
@@ -14,6 +14,7 @@ services:
       - "lbry_rocksdb:/database"
     environment:
       - HUB_COMMAND=scribe
+      - SNAPSHOT_URL=https://snapshots.lbry.com/hub/lbry-rocksdb.zip
     command:
       - "--daemon_url=http://lbry:lbry@127.0.0.1:9245"
       - "--max_query_workers=2"
diff --git a/scripts/entrypoint.sh b/scripts/entrypoint.sh
index 6ff7cb1..bef8856 100755
--- a/scripts/entrypoint.sh
+++ b/scripts/entrypoint.sh
@@ -4,6 +4,22 @@
 
 set -euo pipefail
 
+SNAPSHOT_URL="${SNAPSHOT_URL:-}" #off by default. latest snapshot at https://lbry.com/snapshot/hub
+
+if [[ "$HUB_COMMAND" == "scribe" ]] && [[ -n "$SNAPSHOT_URL" ]] && [[ ! -d /database/lbry-rocksdb ]]; then
+  files="$(ls)"
+  echo "Downloading hub snapshot from $SNAPSHOT_URL"
+  wget --no-verbose --trust-server-names --content-disposition "$SNAPSHOT_URL"
+  echo "Extracting snapshot..."
+  filename="$(grep -vf <(echo "$files") <(ls))" # finds the file that was not there before
+  case "$filename" in
+    *.tgz|*.tar.gz|*.tar.bz2 )  tar xvf "$filename" --directory /database ;;
+    *.zip ) unzip "$filename" -d /database/ ;;
+    * ) echo "Don't know how to extract ${filename}. SNAPSHOT COULD NOT BE LOADED" && exit 1 ;;
+  esac
+  rm "$filename"
+fi
+
 if [ -z "$HUB_COMMAND" ]; then
   echo "HUB_COMMAND env variable must be scribe, herald, or scribe-elastic-sync"
   exit 1