diff options
| author | Kotresh HR <khiremat@redhat.com> | 2019-08-16 15:38:49 +0530 |
|---|---|---|
| committer | Amar Tumballi <amarts@gmail.com> | 2019-08-20 06:20:33 +0000 |
| commit | b1dc7723610325789ba91220ab414e821ec14119 (patch) | |
| tree | 0e37c4db45e4f7f84fd7329e7228bc7c549d07a2 | |
| parent | f138d3fa2237e7fa940ecf17153fd700350c4138 (diff) | |
geo-rep: Fix worker connection issue
All the workers connects to primary slave node. It should
connect to available slave nodes in round robin fashion
and choose different slave node if the corresponding slave
node is down. This patch fixes the same.
Thanks Aravinda for the help in root causing this.
Change-Id: I9f8e7744f4adb8a24833cf173681d109710f98cb
Signed-off-by: Kotresh HR <khiremat@redhat.com>
Updates: bz#1737484
| -rw-r--r-- | geo-replication/syncdaemon/subcmds.py | 3 |
1 files changed, 2 insertions, 1 deletions
diff --git a/geo-replication/syncdaemon/subcmds.py b/geo-replication/syncdaemon/subcmds.py index 4ece7e06b89..8de7db2cdaa 100644 --- a/geo-replication/syncdaemon/subcmds.py +++ b/geo-replication/syncdaemon/subcmds.py @@ -73,7 +73,8 @@ def subcmd_worker(args): Popen.init_errhandler() fcntl.fcntl(args.feedback_fd, fcntl.F_SETFD, fcntl.FD_CLOEXEC) local = GLUSTER("localhost", args.master) - slavehost, slavevol = args.slave.split("::") + slavevol = args.slave.split("::")[-1] + slavehost = args.resource_remote remote = SSH(slavehost, slavevol) remote.connect_remote() local.connect() |