Merge pull request #48 from lbryio/initialize-from-snapshot

Optionally initialize rocksdb from a snapshot

docs/cluster_guide.md

@@ -146,10 +146,12 @@ 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"
       - "--cache_all_tx_hashes"
+      - "--index_address_statuses"
   scribe_elastic_sync:
     image: lbry/hub:${SCRIBE_TAG:-development}
     restart: always
@@ -183,6 +185,7 @@ services:
       - FILTERING_CHANNEL_IDS=770bd7ecba84fd2f7607fb15aedd2b172c2e153f 95e5db68a3101df19763f3a5182e4b12ba393ee8
       - BLOCKING_CHANNEL_IDS=dd687b357950f6f271999971f43c785e8067c3a9 06871aa438032244202840ec59a469b303257cad b4a2528f436eca1bf3bf3e10ff3f98c57bd6c4c6
     command:
+      - "--index_address_statuses"
       - "--daemon_url=http://lbry:lbry@127.0.0.1:9245"
       - "--elastic_host=127.0.0.1"
       - "--elastic_port=9200"

docs/docker_examples/docker-compose.yml

@@ -17,10 +17,12 @@ 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"
       # - "--cache_all_tx_hashes"  # uncomment to keep an index of all tx hashes in memory. This uses lots (10+g) of memory but substantially improves performance.
+      - "--index_address_statuses"
   scribe_elastic_sync:
     depends_on:
       - es01
@@ -41,7 +43,7 @@ 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
@@ -59,6 +61,7 @@ services:
       - FILTERING_CHANNEL_IDS=770bd7ecba84fd2f7607fb15aedd2b172c2e153f 95e5db68a3101df19763f3a5182e4b12ba393ee8
       - BLOCKING_CHANNEL_IDS=dd687b357950f6f271999971f43c785e8067c3a9 06871aa438032244202840ec59a469b303257cad b4a2528f436eca1bf3bf3e10ff3f98c57bd6c4c6
     command:  # for full options, see `herald --help`
+      - "--index_address_statuses"
       - "--daemon_url=http://lbry:lbry@127.0.0.1:9245"
       - "--max_query_workers=4"
       - "--host=0.0.0.0"

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"

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