diff options
| author | Jeff Darcy <jdarcy@redhat.com> | 2017-03-07 18:36:58 -0500 |
|---|---|---|
| committer | Shyamsundar Ranganathan <srangana@redhat.com> | 2017-03-10 14:50:11 -0500 |
| commit | e6c10359ab90178f89aa23ae9328174c2602e88d (patch) | |
| tree | 008b7f03ad551b7eec3e23e29fdd3370c2ac831f | |
| parent | 2579daf23d6039d2b001d6e2392a952b39317af4 (diff) | |
glusterd: don't queue attach reqs before connecting
This was causing USS tests to fail. The underlying problem here is
that if we try to queue the attach request too soon after starting a
brick process then the socket code will get an error trying to write
to the still-unconnected socket. Its response is to shut down the
socket, which causes the queued attach requests to be force-unwound.
There's nothing to retry them, so they effectively never happen and
those bricks (second and succeeding for a snapshot) never become
available.
We *do* have a retry loop for attach requests, but currently break out
as soon as a request is queued - not actually sent. The fix is to
modify that loop so it will wait some more if the rpc connection isn't
even complete yet. Now we break out only when we have a completed
connection *and* a queued request.
Backport of:
> 53e2c875cf97df8337f7ddb5124df2fc6dd37bca
> BUG: 1430148
> Reviewed-on: https://review.gluster.org/16868
Signed-off-by: Jeff Darcy <jdarcy@redhat.com>
BUG: 1431176
Change-Id: Ib6be13646f1fa9072b4a944ab5f13e1b29084841
Reviewed-on: https://review.gluster.org/16887
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>
| -rw-r--r-- | xlators/mgmt/glusterd/src/glusterd-utils.c | 29 |
1 files changed, 18 insertions, 11 deletions
diff --git a/xlators/mgmt/glusterd/src/glusterd-utils.c b/xlators/mgmt/glusterd/src/glusterd-utils.c index c501e1ad00b..97f95adbd11 100644 --- a/xlators/mgmt/glusterd/src/glusterd-utils.c +++ b/xlators/mgmt/glusterd/src/glusterd-utils.c @@ -4913,22 +4913,29 @@ my_callback (struct rpc_req *req, struct iovec *iov, int count, void *v_frame) int send_attach_req (xlator_t *this, struct rpc_clnt *rpc, char *path, int op) { - int ret = -1; - struct iobuf *iobuf = NULL; - struct iobref *iobref = NULL; - struct iovec iov = {0, }; - ssize_t req_size = 0; - call_frame_t *frame = NULL; - gd1_mgmt_brick_op_req brick_req; - void *req = &brick_req; - void *errlbl = &&err; - extern struct rpc_clnt_program gd_brick_prog; + int ret = -1; + struct iobuf *iobuf = NULL; + struct iobref *iobref = NULL; + struct iovec iov = {0, }; + ssize_t req_size = 0; + call_frame_t *frame = NULL; + gd1_mgmt_brick_op_req brick_req; + void *req = &brick_req; + void *errlbl = &&err; + struct rpc_clnt_connection *conn; + extern struct rpc_clnt_program gd_brick_prog; if (!rpc) { gf_log (this->name, GF_LOG_ERROR, "called with null rpc"); return -1; } + conn = &rpc->conn; + if (!conn->connected || conn->disconnected) { + gf_log (this->name, GF_LOG_INFO, "not connected yet"); + return -1; + } + brick_req.op = op; brick_req.name = path; brick_req.input.input_val = NULL; @@ -5046,7 +5053,7 @@ attach_brick (xlator_t *this, (void) build_volfile_path (full_id, path, sizeof(path), NULL); int tries = 0; - while (tries++ <= 10) { + while (tries++ <= 15) { ret = send_attach_req (this, other_brick->rpc, path, GLUSTERD_BRICK_ATTACH); if (!ret) { |