arvindshmicrosoft commented on a change in pull request #270: Support Accumulo
installs on Microsoft Azure
URL: https://github.com/apache/fluo-muchos/pull/270#discussion_r313652677
##########
File path: lib/muchos/config.py
##########
@@ -130,17 +134,30 @@ def force_format(self):
return 'no'
return retval
- def worker_data_dirs(self):
+ def data_dirs_common(self, nodeType):
+ data_dirs = []
+
if self.get_cluster_type() == 'ec2':
- return self.node_type_map()['worker']['mounts']
+ data_dirs = self.node_type_map()[nodeType]['mounts']
elif self.get_cluster_type() == 'existing':
- return self.get('existing', 'data_dirs').split(",")
+ data_dirs = self.get('existing', 'data_dirs').split(",")
+ elif self.get_cluster_type() == 'azure':
+ num_disks = int(self.get("azure","numdisks"))
+ range_var = num_disks + 1
+ for diskNum in range(1, range_var):
+ data_dirs.append(self.get("azure","mount_root") + str(diskNum))
+
+ # we need a minimum of 3 "data directories" in order for the HA
changes to work correctly. If the number of disks specified is < 3, stub the
balance entries.
+ for tmpId in range (len(data_dirs) - 1, 2):
Review comment:
In the process of the Hadoop / ZK HA changes, what we were recommended was
to have separate disks for ZK and Hadoop namenode folders. In this process we
ended up with 3 separate folder locations.
In the case of deployments where the ephemeral disk is the only disk used,
or in case of Azure when there is just 1 data disk attached, we have to "stub"
the extra disk locations. This step does that. I believe post this PR, we will
make the entire Hadoop / ZK HA configuration switchable wherein we will restore
the old behavior if there is a non-HA config selected.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
