From: Vladislav Shpilevoy <v.shpilevoy@tarantool.org> To: Cyrill Gorcunov <gorcunov@gmail.com> Cc: Mons Anderson <v.perepelitsa@corp.mail.ru>, tml <tarantool-patches@dev.tarantool.org> Subject: Re: [Tarantool-patches] [PATCH v4 2/3] cfg: support symbolic evaluation of replication_synchro_quorum Date: Sun, 20 Dec 2020 18:01:25 +0100 [thread overview] Message-ID: <47af00c0-5f40-533c-f51c-84d16c609c11@tarantool.org> (raw) In-Reply-To: <20201218072529.GF14556@grain> >>> static int >>> box_check_replication_synchro_quorum(void) >>> { >>> - int quorum = cfg_geti("replication_synchro_quorum"); >>> + int quorum = 0; >>> + >>> + if (!cfg_isnumber("replication_synchro_quorum")) { >>> + /* >>> + * The formula uses symbolic name 'N' as >>> + * a number of currently registered replicas. >>> + * >>> + * When we're in "checking" mode we should walk >>> + * over all possible number of replicas to make >>> + * sure the formula is correct. >>> + * >>> + * Note that currently VCLOCK_MAX is pretty small >>> + * value but if we gonna increase this limit make >>> + * sure that the cycle won't take too much time. >>> + */ >>> + for (int i = 1; i < VCLOCK_MAX; i++) { >>> + quorum = box_eval_replication_synchro_quorum(i); >>> + if (quorum < 0) >>> + return -1; >>> + } >>> + /* >>> + * Just to make clear the number we return here doesn't >>> + * have any special meaning, only errors are matter. >>> + * The real value is dynamic and will be updated on demand. >>> + */ >> >> 3. Wtf? This function before your patch was supposed to return the >> new quorum value. Like all cfg 'check()' functions. If it can't do that >> now (but it can - just evaluate with the current number of replicas), >> then the function must return only 0 or -1, like all 'binary result' >> functions. Now it simply returns some random number in case the quorum >> is an expression. > > As I pointer in the comment there is no special meaning in the return > value, since we update the quorum on demand. Moreover once we manage > to validate the formula we call box_update_replication_synchro_quorum > which reevaluates the quorum with current number of replicas, thus > to not make same work twice I will return 0 here. I saw the comment and I understood it. But it does not mean the function can now return random values. It does not matter if you return 0 or 1 or whatever else. It still can't be used for anything useful except check < 0. That makes the function result confusing and even useless. Previously quorum value was returned *and used as quorum value*. Now it is used only for < 0 check. Hence, why do you need to return anything except -1 and 0? For example, look at box_check_replication_synchro_timeout(). It returns *timeout value*, which is used to set replication_synchro_timeout. box_set_replication_synchro_quorum() before your patch did the same with box_check_replication_synchro_quorum(). Now it uses check() result only to compare it with < 0, which is *confusing* - why would it need to return anything but -1 and 0 then? Here is what I want to see: ==================== diff --git a/src/box/box.cc b/src/box/box.cc index ff5c27743..751ec4733 100644 --- a/src/box/box.cc +++ b/src/box/box.cc @@ -640,8 +640,6 @@ box_eval_replication_synchro_quorum(int nr_replicas) static int box_check_replication_synchro_quorum(void) { - int quorum = 0; - if (!cfg_isnumber("replication_synchro_quorum")) { /* * The formula uses symbolic name 'N' as @@ -656,8 +654,7 @@ box_check_replication_synchro_quorum(void) * sure that the cycle won't take too much time. */ for (int i = 1; i < VCLOCK_MAX; i++) { - quorum = box_eval_replication_synchro_quorum(i); - if (quorum < 0) + if (box_eval_replication_synchro_quorum(i) < 0) return -1; } /* @@ -666,17 +663,15 @@ box_check_replication_synchro_quorum(void) * The real value is dynamic and will be updated on demand. */ return 0; - } else { - quorum = cfg_geti("replication_synchro_quorum"); } - + int quorum = cfg_geti("replication_synchro_quorum"); if (quorum <= 0 || quorum >= VCLOCK_MAX) { diag_set(ClientError, ER_CFG, "replication_synchro_quorum", "the value must be greater than zero and less than " "maximal number of replicas"); return -1; } - return quorum; + return 0; } static double @@ -877,7 +872,7 @@ box_check_config(void) box_check_replication_connect_timeout(); box_check_replication_connect_quorum(); box_check_replication_sync_lag(); - if (box_check_replication_synchro_quorum() < 0) + if (box_check_replication_synchro_quorum() != 0) diag_raise(); if (box_check_replication_synchro_timeout() < 0) diag_raise(); @@ -1057,8 +1052,7 @@ box_update_replication_synchro_quorum(void) int box_set_replication_synchro_quorum(void) { - int value = box_check_replication_synchro_quorum(); - if (value < 0) + if (box_check_replication_synchro_quorum() != 0) return -1; box_update_replication_synchro_quorum(); return 0;
next prev parent reply other threads:[~2020-12-20 17:01 UTC|newest] Thread overview: 20+ messages / expand[flat|nested] mbox.gz Atom feed top 2020-12-14 11:39 [Tarantool-patches] [PATCH v4 0/3] qsync: evaluate replication_synchro_quorum dynamically Cyrill Gorcunov 2020-12-14 11:39 ` [Tarantool-patches] [PATCH v4 1/3] cfg: add cfg_isnumber helper Cyrill Gorcunov 2020-12-14 11:39 ` [Tarantool-patches] [PATCH v4 2/3] cfg: support symbolic evaluation of replication_synchro_quorum Cyrill Gorcunov 2020-12-16 13:21 ` Serge Petrenko 2020-12-16 13:35 ` Cyrill Gorcunov 2020-12-17 23:17 ` Vladislav Shpilevoy 2020-12-18 7:25 ` Cyrill Gorcunov 2020-12-20 17:01 ` Vladislav Shpilevoy [this message] 2020-12-20 18:28 ` Cyrill Gorcunov 2020-12-21 17:48 ` Vladislav Shpilevoy 2020-12-21 17:49 ` Vladislav Shpilevoy 2020-12-21 20:02 ` Cyrill Gorcunov 2020-12-21 20:12 ` Cyrill Gorcunov 2020-12-14 11:39 ` [Tarantool-patches] [PATCH v4 3/3] test: add replication/gh-5446-qsync-eval-quorum.test.lua Cyrill Gorcunov 2020-12-16 13:25 ` Serge Petrenko 2020-12-17 23:18 ` Vladislav Shpilevoy 2020-12-18 8:14 ` Cyrill Gorcunov 2020-12-20 17:01 ` Vladislav Shpilevoy 2020-12-20 18:27 ` Cyrill Gorcunov 2020-12-21 16:05 ` Cyrill Gorcunov
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=47af00c0-5f40-533c-f51c-84d16c609c11@tarantool.org \ --to=v.shpilevoy@tarantool.org \ --cc=gorcunov@gmail.com \ --cc=tarantool-patches@dev.tarantool.org \ --cc=v.perepelitsa@corp.mail.ru \ --subject='Re: [Tarantool-patches] [PATCH v4 2/3] cfg: support symbolic evaluation of replication_synchro_quorum' \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: link
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox