From: Vladimir Davydov <vdavydov.dev@gmail.com> To: Ilya Kosarev <i.kosarev@tarantool.org> Cc: tarantool-patches@freelists.org, georgy@tarantool.org, i.kosarev@corp.mail.ru Subject: Re: [PATCH v3] iproto: report active connections number Date: Wed, 27 Feb 2019 11:50:17 +0300 [thread overview] Message-ID: <20190227085017.xscldynpdvpmgmwx@esperanza> (raw) In-Reply-To: <20190226180420.3129-1-i.kosarev@tarantool.org> On Tue, Feb 26, 2019 at 09:04:20PM +0300, Ilya Kosarev wrote: > Now there is new member in box.stat.net() called "CONNECTIONS" > which is a number of active iproto connections. > > Closes #3905 > --- > https://github.com/tarantool/tarantool/tree/i.kosarev/gh-3905-report-connections-number > https://github.com/tarantool/tarantool/issues/3905 > > Changes in v3: > - CONNECTIONS counter is not rmeanish anymore > - fiber.sleep replaced with test_run:wait_cond in test > > src/box/iproto.cc | 3 ++- > src/box/lua/stat.c | 44 ++++++++++++++++++++++++++++++++++++++ > test/box/stat_net.result | 37 ++++++++++++++++++++++++++++++++ > test/box/stat_net.test.lua | 13 +++++++++++ > 4 files changed, 96 insertions(+), 1 deletion(-) > > diff --git a/src/box/iproto.cc b/src/box/iproto.cc > index 863eb4f06..1c506d9c8 100644 > --- a/src/box/iproto.cc > +++ b/src/box/iproto.cc > @@ -271,6 +271,7 @@ enum rmean_net_name { > }; > > const char *rmean_net_strings[IPROTO_LAST] = { "SENT", "RECEIVED" }; > +const char *conn_n_name = "CONNECTIONS"; What's the point in having this string defined in iproto.cc? It's only used in lua/stat.c. > > static void > tx_process_destroy(struct cmsg *m); > @@ -475,7 +476,7 @@ struct iproto_connection > char salt[IPROTO_SALT_SIZE]; > }; > > -static struct mempool iproto_connection_pool; > +struct mempool iproto_connection_pool; > static RLIST_HEAD(stopped_connections); > > /** > diff --git a/src/box/lua/stat.c b/src/box/lua/stat.c > index 3fce81f61..cedfaba8a 100644 > --- a/src/box/lua/stat.c > +++ b/src/box/lua/stat.c > @@ -45,12 +45,15 @@ > #include <info.h> > #include "lua/info.h" > #include "lua/utils.h" > +#include "small/mempool.h" > > extern struct rmean *rmean_box; > extern struct rmean *rmean_error; > /** network statistics (iproto & cbus) */ > extern struct rmean *rmean_net; > extern struct rmean *rmean_tx_wal_bus; > +extern struct mempool iproto_connection_pool; No, please don't export iproto_connection_pool. Instead, add a helper function to report the connection count, similar to iproto_mem_used(). > +extern const char *conn_n_name; > > static void > fill_stat_item(struct lua_State *L, int rps, int64_t total) > @@ -79,6 +82,29 @@ set_stat_item(const char *name, int rps, int64_t total, void *cb_ctx) > return 0; > } > > +static void > +fill_conn_n(struct lua_State *L) > +{ > + lua_pushstring(L, "total"); Why not simply print "CONNECTIONS: 10"? I don't see any point in the "total" field. > + lua_pushnumber(L, mempool_count(&iproto_connection_pool)); > + lua_settable(L, -3); > +} > + > +static int > +set_conn_n(void *cb_ctx) > +{ > + struct lua_State *L = (struct lua_State *) cb_ctx; > + > + lua_pushstring(L, conn_n_name); > + lua_newtable(L); > + > + fill_conn_n(L); > + > + lua_settable(L, -3); > + > + return 0; > +} > + > /** > * A stat_foreach() callback used to handle access to e.g. > * box.stats.DELETE. > @@ -96,6 +122,19 @@ seek_stat_item(const char *name, int rps, int64_t total, void *cb_ctx) > return 1; > } > > +static int > +seek_conn_n(void *cb_ctx) > +{ > + struct lua_State *L = (struct lua_State *) cb_ctx; > + if (strcmp(conn_n_name, lua_tostring(L, -1)) != 0) > + return 0; > + > + lua_newtable(L); > + fill_conn_n(L); > + > + return 1; > +} > + What's the point in making these functions look similar to seek_stat_item/set_stat_item? They aren't callbacks. I don't think we even need them - they're trivial so we could inline them without hurting readability. > static int > lbox_stat_index(struct lua_State *L) > { > @@ -140,6 +179,9 @@ static int > lbox_stat_net_index(struct lua_State *L) > { > luaL_checkstring(L, -1); > + int res = seek_conn_n(L); > + if (res) > + return res; > return rmean_foreach(rmean_net, seek_stat_item, L); > } > > @@ -148,6 +190,8 @@ lbox_stat_net_call(struct lua_State *L) > { > lua_newtable(L); > rmean_foreach(rmean_net, set_stat_item, L); > + set_conn_n(L); > + > return 1; > }
prev parent reply other threads:[~2019-02-27 8:50 UTC|newest] Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top 2019-02-26 18:04 Ilya Kosarev 2019-02-27 8:50 ` Vladimir Davydov [this message]
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=20190227085017.xscldynpdvpmgmwx@esperanza \ --to=vdavydov.dev@gmail.com \ --cc=georgy@tarantool.org \ --cc=i.kosarev@corp.mail.ru \ --cc=i.kosarev@tarantool.org \ --cc=tarantool-patches@freelists.org \ --subject='Re: [PATCH v3] iproto: report active connections number' \ /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