From a60fc2ddc03134fb23c5ed5c0bcb195e1649416b Mon Sep 17 00:00:00 2001
From: Sanju Rakonde <srakonde@redhat.com>
Date: Wed, 21 Feb 2018 12:46:25 +0530
Subject: glusterd: handling brick termination in brick-mux

Problem: There's a race between the last glusterfs_handle_terminate()
response sent to glusterd and the kill that happens immediately if the
terminated brick is the last brick.

Solution: When it is a last brick for the brick process, instead of glusterfsd
killing itself, glusterd will kill the process in case of brick multiplexing.
And also changing gf_attach utility accordingly.

Change-Id: I386c19ca592536daa71294a13d9fc89a26d7e8c0
fixes: bz#1545048
BUG: 1545048
Signed-off-by: Sanju Rakonde <srakonde@redhat.com>
---
 glusterfsd/src/gf_attach.c | 41 ++++++++++++++++++++++++++++++++++++++---
 1 file changed, 38 insertions(+), 3 deletions(-)

(limited to 'glusterfsd/src/gf_attach.c')

diff --git a/glusterfsd/src/gf_attach.c b/glusterfsd/src/gf_attach.c
index 3f248292ddf..0eb4868263b 100644
--- a/glusterfsd/src/gf_attach.c
+++ b/glusterfsd/src/gf_attach.c
@@ -11,6 +11,9 @@
 #include <stdio.h>
 #include <stdlib.h>
 #include <unistd.h>
+#include <sys/types.h>
+#include <sys/wait.h>
+#include <signal.h>
 
 //#include "config.h"
 #include "glusterfs.h"
@@ -23,6 +26,7 @@
 
 int done = 0;
 int rpc_status;
+glfs_t *fs;
 
 struct rpc_clnt_procedure gf_attach_actors[GLUSTERD_BRICK_MAXVALUE] = {
         [GLUSTERD_BRICK_NULL] = {"NULL", NULL },
@@ -71,11 +75,43 @@ my_notify (struct rpc_clnt *rpc, void *mydata,
 }
 
 int32_t
-my_callback (struct rpc_req *req, struct iovec *iov, int count, void *frame)
+my_callback (struct rpc_req *req, struct iovec *iov, int count, void *v_frame)
 {
+        gd1_mgmt_brick_op_rsp     rsp;
+        dict_t          *dict   = NULL;
+        pid_t            pid    = -1;
+        int              ret    = -1;
+        xlator_t        *this   = NULL;
+
+        this = fs->ctx->master;
+        memset (&rsp, 0, sizeof (rsp));
+
+        ret = xdr_to_generic (*iov, &rsp, (xdrproc_t)xdr_gd1_mgmt_brick_op_rsp);
+
+        if (ret < 0) {
+                fprintf (stderr, "xdr decoding failed\n");
+                goto out;
+        }
+        GF_PROTOCOL_DICT_UNSERIALIZE (this, dict,
+                                      (rsp.output.output_val),
+                                      (rsp.output.output_len),
+                                      ret, rsp.op_errno, out);
+        if (dict) {
+                if (dict_get_int32 (dict, "last_brick_terminated", &pid) == 0) {
+                        int status = 0;
+
+                        gf_log ("gf_attach", GF_LOG_INFO, "Killing %d", pid);
+                        kill (pid, SIGTERM);
+                        waitpid (pid, &status, 0);
+                }
+                dict_unref (dict);
+        }
+
         rpc_status = req->rpc_status;
         done = 1;
-        return 0;
+        ret = 0;
+out:
+        return ret;
 }
 
 /* copied from gd_syncop_submit_request */
@@ -170,7 +206,6 @@ usage (char *prog)
 int
 main (int argc, char *argv[])
 {
-        glfs_t                  *fs;
         struct rpc_clnt         *rpc;
         dict_t                  *options;
         int                     ret;
-- 
cgit