diff options
author | Jeff Darcy <jdarcy@redhat.com> | 2017-02-01 22:00:32 -0500 |
---|---|---|
committer | Shyamsundar Ranganathan <srangana@redhat.com> | 2017-02-02 19:43:24 -0500 |
commit | 0a0112b2c02a30bcb7eca8fa9ecb7fbbe84aa7f8 (patch) | |
tree | f65dc68856f72a362175a9eb4ec14533ea085d93 /rpc | |
parent | ade302a5141a95ec1a4d6de5a98f745613ed1446 (diff) |
socket: retry connect immediately if it fails
Previously we relied on a complex dance of setting flags, shutting
down the socket, tearing stuff down, getting an event, tearing more
stuff down, and waiting for a higher-level retry. What we really
need, in the case where we're just trying to connect prematurely e.g.
to a brick that hasn't fully come up yet, is a simple retry of the
connect(2) call.
This was discovered by observing failures in ec-new-entry.t with
multiplexing enabled, but probably fixes other random failures as
well.
Backport of:
> Change-Id: Ibedb8942060bccc96b02272a333c3002c9b77d4c
> BUG: 1385758
> Reviewed-on: https://review.gluster.org/16510
BUG: 1418091
Change-Id: I4bac26929a12cabcee4f9e557c8b4d520948378b
Signed-off-by: Jeff Darcy <jdarcy@redhat.com>
Reviewed-on: https://review.gluster.org/16533
Smoke: Gluster Build System <jenkins@build.gluster.org>
NetBSD-regression: NetBSD Build System <jenkins@build.gluster.org>
CentOS-regression: Gluster Build System <jenkins@build.gluster.org>
Reviewed-by: Shyamsundar Ranganathan <srangana@redhat.com>
Diffstat (limited to 'rpc')
-rw-r--r-- | rpc/rpc-transport/socket/src/socket.c | 38 |
1 files changed, 36 insertions, 2 deletions
diff --git a/rpc/rpc-transport/socket/src/socket.c b/rpc/rpc-transport/socket/src/socket.c index b128be30603..d2649e6306e 100644 --- a/rpc/rpc-transport/socket/src/socket.c +++ b/rpc/rpc-transport/socket/src/socket.c @@ -2944,6 +2944,33 @@ socket_fix_ssl_opts (rpc_transport_t *this, socket_private_t *priv, } } +/* + * If we might just be trying to connect prematurely, e.g. to a brick that's + * slow coming up, all we need is a simple retry. Don't worry about sleeping + * in some arbitrary thread. The connect(2) could already have the exact same + * effect, and we deal with it in that case so we can deal with it for sleep(2) + * as well. + */ +static int +connect_loop (int sockfd, const struct sockaddr *addr, socklen_t addrlen) +{ + int ret; + int connect_fails = 0; + + for (;;) { + ret = connect (sockfd, addr, addrlen); + if (ret >= 0) { + break; + } + if ((errno != ENOENT) || (++connect_fails >= 5)) { + break; + } + sleep (1); + } + + return ret; +} + static int socket_connect (rpc_transport_t *this, int port) { @@ -3105,8 +3132,15 @@ socket_connect (rpc_transport_t *this, int port) } } - ret = connect (priv->sock, SA (&this->peerinfo.sockaddr), - this->peerinfo.sockaddr_len); + if (ign_enoent) { + ret = connect_loop (priv->sock, + SA (&this->peerinfo.sockaddr), + this->peerinfo.sockaddr_len); + } else { + ret = connect (priv->sock, + SA (&this->peerinfo.sockaddr), + this->peerinfo.sockaddr_len); + } if (ret == -1 && errno == ENOENT && ign_enoent) { gf_log (this->name, GF_LOG_WARNING, |