YUKI Hiroshi
null+****@clear*****
Fri Nov 21 20:29:15 JST 2014
YUKI Hiroshi 2014-11-21 20:29:15 +0900 (Fri, 21 Nov 2014) New Revision: 3739f10b0e531042de1f8f312d10897798c5f6ab https://github.com/droonga/droonga.org/commit/3739f10b0e531042de1f8f312d10897798c5f6ab Message: Add note to specify maximum number of inpouring messages per second Modified files: _po/ja/roadmap/index.po _po/ja/tutorial/1.0.8/benchmark/index.po ja/roadmap/index.md ja/tutorial/1.0.8/benchmark/index.md tutorial/1.0.8/benchmark/index.md Modified: _po/ja/roadmap/index.po (+19 -6) =================================================================== --- _po/ja/roadmap/index.po 2014-11-20 18:37:11 +0900 (e13c201) +++ _po/ja/roadmap/index.po 2014-11-21 20:29:15 +0900 (9884a9c) @@ -1,7 +1,7 @@ msgid "" msgstr "" "Project-Id-Version: PACKAGE VERSION\n" -"PO-Revision-Date: 2014-11-02 21:27+0900\n" +"PO-Revision-Date: 2014-03-05 12:52+0900\n" "Language: ja\n" "MIME-Version: 1.0\n" "Content-Type: text/plain; charset=UTF-8\n" @@ -30,16 +30,17 @@ msgstr "" msgid "## 1.0.8" msgstr "" -msgid " * Make drop-in replacement of log search system based on Groonga." +msgid "" +" * Make drop-in replacement of log search system based on Groonga.\n" +" * Optimize performance of `search` for unsharded replicas.\n" +" (Currently, mechanisms of hash-based sharding always affect for replicas.\n" +" They should be skipped for unsharded replicas.)" msgstr "" msgid "## 1.x.x" msgstr "" msgid "" -" * Optimize performance of `search` for unsharded replicas.\n" -" (Currently, mechanisms of hash-based sharding always affect for replicas.\n" -" They should be skipped for unsharded replicas.)\n" " * Support hash-based sharding of replicas, by `droonga-enigine-join` and `dro" "onga-engine-unjoin` commands.\n" " * Support non-stop modification of cluster structure.\n" @@ -84,13 +85,19 @@ msgstr "" msgid "# `droonga-http-server`" msgstr "" +msgid "## 1.0.9" +msgstr "" + msgid "" " * Detect available `droonga-engine` nodes automatically.\n" " (Nodes should fetch `catalog.json` from initially connected `droonga-engine" "` node via the `catalog.fetch` command and refresh connections based on the cl" "uster definition.\n" " * Prevent to send request messages to dead `droonga-engine` nodes.\n" -" (Nodes should join to the orchestration based on Serf.)\n" +" (Nodes should join to the orchestration based on Serf.)" +msgstr "" + +msgid "" " * Better dashboard.\n" " (For example, search columns should be listed and filtered based on the sch" "ema information.\n" @@ -101,5 +108,11 @@ msgstr "" msgid "# `express-droonga`" msgstr "" +msgid "## 1.0.7" +msgstr "" + msgid " * Make connections to `droonga-engine` nodes redundant." msgstr "" + +msgid " * ???" +msgstr "" Modified: _po/ja/tutorial/1.0.8/benchmark/index.po (+23 -3) =================================================================== --- _po/ja/tutorial/1.0.8/benchmark/index.po 2014-11-20 18:37:11 +0900 (6151a2d) +++ _po/ja/tutorial/1.0.8/benchmark/index.po 2014-11-21 20:29:15 +0900 (166dc4b) @@ -608,6 +608,7 @@ msgid "" " droonga-send --server=node0 \\\n" " --server=node1 \\\n" " --server=node2 \\\n" +" --messages-per-second=100 \\\n" " --report-throughput)\n" "~~~" msgstr "" @@ -623,13 +624,32 @@ msgstr "" "流し込むと、データベースが壊れてしまいます。" msgid "" +"To reduce traffic and system load, you should specify maximum number of inpour" +"ing messages per second by the `--messages-per-second` option.\n" +"If too many messages rush into the Droonga cluster, they may overflow - Droong" +"a may eat up the RAM and slow down the system." +msgstr "" +"トラフィックとシステムの負荷を軽減するために、1秒あたりに流入するメッセージの量を`--messages-per-second`オプションで制限するようにし" +"てください。\n" +"大量のメッセージが一度にDroongaクラスタに流れ込むと、システムの限界を超えてしまい、Droongaがメモリを食い潰して、システムを非常に低速にしてしま" +"う恐れがあります。" + +msgid "" "This may take much time.\n" +"For example, with the option `--messages-per-second=100`, 1500000 records will" +" be synchronized in about 4 hours (we can estimate the required time like: `15" +"0000 / 100 / 60 / 60`)." +msgstr "" +"この操作にも時間がかかります。\n" +"例えば `--messages-per-second=100` と指定した場合、150万件のレコードを同期するにはだいたい4時間ほどかかります。（必要な時間" +"は `150000 / 100 / 60 / 60` のような計算式で見積もれます）" + +msgid "" "After all, now you have two HTTP servers: Groonga HTTP server with the port `1" "0041`, and Droonga HTTP Servers with the port `10042`." msgstr "" -"この操作にも時間がかかります。\n" -"それが完了したら、`10041`ポートを監視するGroonga HTTPサーバと、`10042`ポートを監視するDroonga HTTPサーバの、2つのHT" -"TPサーバがある状態となります。" +"以上の手順により、`10041`ポートを監視するGroonga HTTPサーバと、`10042`ポートを監視するDroonga HTTPサーバの、2つのHT" +"TPサーバを用意できます。" msgid "### Set up the client" msgstr "### クライアントをセットアップする" Modified: ja/roadmap/index.md (+12 -5) =================================================================== --- ja/roadmap/index.md 2014-11-20 18:37:11 +0900 (b02bd9e) +++ ja/roadmap/index.md 2014-11-21 20:29:15 +0900 (19808fa) @@ -20,12 +20,12 @@ layout: ja ## 1.0.8 * Make drop-in replacement of log search system based on Groonga. - -## 1.x.x - * Optimize performance of `search` for unsharded replicas. (Currently, mechanisms of hash-based sharding always affect for replicas. They should be skipped for unsharded replicas.) + +## 1.x.x + * Support hash-based sharding of replicas, by `droonga-enigine-join` and `droonga-engine-unjoin` commands. * Support non-stop modification of cluster structure. (Currently, we have to stop updating data while modifying cluster structure. @@ -53,19 +53,26 @@ layout: ja # `droonga-http-server` -## 1.x.x +## 1.0.9 * Detect available `droonga-engine` nodes automatically. (Nodes should fetch `catalog.json` from initially connected `droonga-engine` node via the `catalog.fetch` command and refresh connections based on the cluster definition. * Prevent to send request messages to dead `droonga-engine` nodes. (Nodes should join to the orchestration based on Serf.) + +## 1.x.x + * Better dashboard. (For example, search columns should be listed and filtered based on the schema information. This better dashboard will be developed in a separate repository to be shared by Groonga and Droogna...) # `express-droonga` -## 1.x.x +## 1.0.7 * Make connections to `droonga-engine` nodes redundant. +## 1.x.x + + * ??? + Modified: ja/tutorial/1.0.8/benchmark/index.md (+7 -1) =================================================================== --- ja/tutorial/1.0.8/benchmark/index.md 2014-11-20 18:37:11 +0900 (9eb49ff) +++ ja/tutorial/1.0.8/benchmark/index.md 2014-11-21 20:29:15 +0900 (687df47) @@ -314,14 +314,20 @@ DroongaのHTTPサーバが動作しており、`10042`番のポートを監視 droonga-send --server=node0 \ --server=node1 \ --server=node2 \ + --messages-per-second=100 \ --report-throughput) ~~~ スキーマ定義とインデックスの定義については単一のエンドポイントに送るように注意して下さい。 Droongaは複数のノードに並行してバラバラに送られたスキーマ変更コマンドをソートすることができないので、スキーマ定義のリクエストを複数のエンドポイントに流し込むと、データベースが壊れてしまいます。 +トラフィックとシステムの負荷を軽減するために、1秒あたりに流入するメッセージの量を`--messages-per-second`オプションで制限するようにしてください。 +大量のメッセージが一度にDroongaクラスタに流れ込むと、システムの限界を超えてしまい、Droongaがメモリを食い潰して、システムを非常に低速にしてしまう恐れがあります。 + この操作にも時間がかかります。 -それが完了したら、`10041`ポートを監視するGroonga HTTPサーバと、`10042`ポートを監視するDroonga HTTPサーバの、2つのHTTPサーバがある状態となります。 +例えば `--messages-per-second=100` と指定した場合、150万件のレコードを同期するにはだいたい4時間ほどかかります。（必要な時間は `150000 / 100 / 60 / 60` のような計算式で見積もれます） + +以上の手順により、`10041`ポートを監視するGroonga HTTPサーバと、`10042`ポートを監視するDroonga HTTPサーバの、2つのHTTPサーバを用意できます。 ### クライアントをセットアップする Modified: tutorial/1.0.8/benchmark/index.md (+6 -0) =================================================================== --- tutorial/1.0.8/benchmark/index.md 2014-11-20 18:37:11 +0900 (bca1bfb) +++ tutorial/1.0.8/benchmark/index.md 2014-11-21 20:29:15 +0900 (1e524f7) @@ -305,13 +305,19 @@ Then dump schemas and data separately and load them to the Droonga cluster. droonga-send --server=node0 \ --server=node1 \ --server=node2 \ + --messages-per-second=100 \ --report-throughput) ~~~ Note that you must send requests for schema and indexes to just one endpoint. Parallel sending of schema definition requests for multiple endpoints will break the database, because Droonga cannot sort schema changing commands sent to each node in parallel. +To reduce traffic and system load, you should specify maximum number of inpouring messages per second by the `--messages-per-second` option. +If too many messages rush into the Droonga cluster, they may overflow - Droonga may eat up the RAM and slow down the system. + This may take much time. +For example, with the option `--messages-per-second=100`, 1500000 records will be synchronized in about 4 hours (we can estimate the required time like: `150000 / 100 / 60 / 60`). + After all, now you have two HTTP servers: Groonga HTTP server with the port `10041`, and Droonga HTTP Servers with the port `10042`. -------------- next part -------------- HTML����������������������������...下載
