[MB-5499] data path is reset when node is rebalanced out so 1.8.1behavior will be different when node is added back to the cluster Created: 07/Jun/12  Updated: 09/Jan/13  Resolved: 04/Jul/12

Status: Closed
Project: Couchbase Server
Component/s: documentation, ns_server
Affects Version/s: 1.8.1-release-candidate
Fix Version/s: 1.8.1
Security Level: Public

Type: Bug Priority: Blocker
Reporter: Karan Kumar (Inactive) Assignee: MC Brown (Inactive)
Resolution: Fixed Votes: 0
Labels: 1.8.1-release-notes
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified
Environment: 181-884-rel

Attachments: GZip Archive 10.1.3.114-8091-diag.txt.gz     GZip Archive 10.1.3.115-8091-diag.txt.gz    

 Description   
To reproduce this:-
1) Set data_path of 2 nodes as /mnt
storage path after fresh installation of nodes:-
storage":{"ssd":[],"hdd":[{"path":"/mnt","quotaMb":"none","state":"ok"}]}

2) Rebalance in a node
after rebalancing in:-
"storage":{"ssd":[],"hdd":[{"path":"/mnt","quotaMb":"none","state":"ok"}]}

3) Rebalance out a node
after rebalancing the node out
"storage":{"ssd":[],"hdd":[{"path":"/opt/couchbase/var/lib/couchbase/data"}]]

4) Rebalance in the same node that is removed
after rebalancing the node out
"storage":{"ssd":[],"hdd":[{"path":"/opt/couchbase/var/lib/couchbase/data"}]]

The storage path should get updated to /mnt instead of the default path.

This is a regression from 180, cross checked this with 180 build.

 Comments   
Comment by Karan Kumar (Inactive) [ 07/Jun/12 ]
Conflicting configuration changes to field {node,'ns_1@10.1.3.115',memcached}:
[{'_vclock',[{'ns_1@10.1.3.115',{2,63506323521}}]},
{dedicated_port,11209},
{bucket_engine,"/opt/couchbase/lib/memcached/bucket_engine.so"},
{engines,
[{membase,
[{engine,"/opt/couchbase/lib/memcached/ep.so"},
{initfile,"/opt/couchbase/etc/couchbase/init.sql"},
{static_config_string,
"vb0=false;waitforwarmup=false;failpartialwarmup=false;shardpattern=%d/%b-%i.mb;db_strategy=multiMTVBDB"}]},
{memcached,
[{engine,"/opt/couchbase/lib/memcached/default_engine.so"},
{static_config_string,"vb0=true"}]}]},
{port,11210},
{dedicated_port,11209},
{dbdir,"/opt/couchbase/var/lib/couchbase/data"},
{admin_user,"_admin"},
{admin_pass,"_admin"},
{verbosity,[]}] and
[{'_vclock',
[{'ns_1@10.1.3.115',{1,63506323509}},{'ns_1@127.0.0.1',{3,63506323463}}]},
{dbdir,"/mnt"},
{dedicated_port,11209},
{bucket_engine,"/opt/couchbase/lib/memcached/bucket_engine.so"},
{engines,
[{membase,
[{engine,"/opt/couchbase/lib/memcached/ep.so"},
{initfile,"/opt/couchbase/etc/couchbase/init.sql"},
{static_config_string,
"vb0=false;waitforwarmup=false;failpartialwarmup=false;shardpattern=%d/%b-%i.mb;db_strategy=multiMTVBDB"}]},
{memcached,
[{engine,"/opt/couchbase/lib/memcached/default_engine.so"},
{static_config_string,"vb0=true"}]}]},
{port,11210},
{dedicated_port,11209},
{admin_user,"_admin"},
{admin_pass,"_admin"},
{verbosity,[]}], choosing the former, which looks newer.
Comment by Farshid Ghods (Inactive) [ 07/Jun/12 ]
Aliaksey,

if this is the new expected behavior we need to document this and notify our users that every time the is rebalanced in/out they have to re-initialze the node otherwise things like data_path or other setting will get reset
Comment by Karan Kumar (Inactive) [ 07/Jun/12 ]
Not a bug on 181.

We will have to update all the test in testrunner to do the same.
Comment by Aleksey Kondratenko [ 07/Jun/12 ]
This is old expected behavior. But 1.8.0 (and some releases before that) got it wrong
Comment by Karan Kumar (Inactive) [ 14/Jun/12 ]
Assigning this to MC to add this to the release notes.

When nodes are rebalanced out from 181 cluster, their are going to reset to default data path, users have to reinitialize it, before adding it back.
Comment by MC Brown (Inactive) [ 04/Jul/12 ]
Documentation updated and the release notes written.
Generated at Mon Apr 21 03:22:49 CDT 2014 using JIRA 5.2.4#845-sha1:c9f4cc41abe72fb236945343a1f485c2c844dac9.