[JCBC-28] refactor the entire cluster stream connection Created: 03/Apr/12  Updated: 06/Sep/13

Status: Reopened
Project: Couchbase Java Client
Component/s: Core
Affects Version/s: None
Fix Version/s: .future
Security Level: Public

Type: Improvement Priority: Major
Reporter: Matt Ingenthron Assignee: Michael Nitschinger
Resolution: Unresolved Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified


 Description   
Because of the codebase's legacy, the handling of the Bucket and Configuration is rather odd. It used to exist outside the client to serve a different purpose. At that time, not changing the client internals was desirable.

Fast forwarding to now, the internals should be updated to have the NodeLocator or the connection abstract away much of the configuration details.




[JCBC-18] NPE if hostnames in server bootstrap list are mixed case Created: 12/Mar/12  Updated: 19/Jul/13

Status: Reopened
Project: Couchbase Java Client
Component/s: Core
Affects Version/s: 1.1-dp4
Fix Version/s: .future
Security Level: Public

Type: Bug Priority: Major
Reporter: Matt Ingenthron Assignee: Michael Nitschinger
Resolution: Unresolved Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified


 Description   
A user described a scenario where using mixed case in their URIs lead to an NPE. This is from the map lookup, since what the couchbase cluster sends us is different than what the user entered, I think.

See: http://www.couchbase.com/forums/thread/java-client-101-exception-using-couchbaseclient-servlet-filter




[JCBC-15] add showtype-options to documentation Created: 23/Feb/12  Updated: 06/Sep/13

Status: Reopened
Project: Couchbase Java Client
Component/s: Documentation
Affects Version/s: None
Fix Version/s: .future
Security Level: Public

Type: Improvement Priority: Major
Reporter: Matt Ingenthron Assignee: Michael Nitschinger
Resolution: Unresolved Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified


 Description   
Many of the Java docs should show the type in the options summary.

 Comments   
Comment by Michael Nitschinger [ 15/Nov/12 ]
Can you give me a quick example on what you mean? Reassign it back to me then and I'll fix it!
Comment by Matt Ingenthron [ 15/Nov/12 ]
If you look at the docs, there are many places where we have types that are returned, but we don't sufficiently describe those types. For example:
http://www.couchbase.com/docs/couchbase-sdk-java-1.1/couchbase-sdk-java-set-add.html#table-couchbase-sdk_java_add

It mentions the OperationFuture, but nowhere really tell how to use it (to my knowledge).

You should be able to work with MC on the right way to fix these.




[JCBC-117] mention that OperationFuture.get(tmo) changes state when timeout has been reached Created: 24/Sep/12  Updated: 06/Sep/13

Status: Reopened
Project: Couchbase Java Client
Component/s: Documentation
Affects Version/s: 1.0.3
Fix Version/s: .future
Security Level: Public

Type: Improvement Priority: Major
Reporter: Mark Nunberg Assignee: Michael Nitschinger
Resolution: Unresolved Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Issue Links:
Dependency
blocks JCBC-114 Command Futures never receive results... Reopened

 Description   
get(tmo) should not change the underlying state of the command to being timed out. It should simply respond with a TimeoutException but allow the command to continue.

Specifically, when the arg-tmo (timeout passed as an argument) expires, the underlying command is marked as timed out. For example, if one waits for 50ms on the command and a response has not been received within that time, the command is now dead ('TIMEDOUT', or similar) and waiting again will not help.

It is understandable that some code might rely on the old behavior, so at the very least, this should be documented in 'BIG RED LETTERS' in the get(tmo) method.

 Comments   
Comment by Matt Ingenthron [ 05/Oct/12 ]
Please explain further.
Comment by Michael Nitschinger [ 16/Oct/12 ]
Hey Mark,

Can you explain in more detail what you want to see changed? When the argument is timed-out what should happen then with it?

Thanks,
Michael
Comment by Matt Ingenthron [ 24/Oct/12 ]
As currently designed, the client uses get() to determine timeout. This is not going to change at the moment. There's no other appropriate place internal to the client to check for this timeout of the operation at the moment.
Comment by Mark Nunberg [ 24/Oct/12 ]
Moving this as a documentation bug
Comment by Matt Ingenthron [ 06/Nov/12 ]
Michael, I'd like you to give this one a shot as your first docs bug, I'll help you with this as needed.




[JCBC-114] Command Futures never receive results after rebalance-out (or other sorts of topology/network changes) Created: 17/Sep/12  Updated: 06/Sep/13

Status: Reopened
Project: Couchbase Java Client
Component/s: Documentation
Affects Version/s: 1.0.3
Fix Version/s: .future
Security Level: Public

Type: Bug Priority: Major
Reporter: Mark Nunberg Assignee: Michael Nitschinger
Resolution: Unresolved Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Issue Links:
Dependency
depends on JCBC-117 mention that OperationFuture.get(tmo)... Reopened

 Comments   
Comment by Mark Nunberg [ 03/Oct/12 ]
This is a real blocker, and seems to be related to a few vbuckets. This issue is preventing me from properly measuring command durations
Comment by Farshid Ghods (Inactive) [ 03/Oct/12 ]
Matt/Rags,

This issue is a blocker for executing more integration tests on java sdk. are there workarounds to avoid this use case or a fix on the way ?
Please assign this back to Mark if more information or logs needed for this issue
Comment by Matt Ingenthron [ 04/Oct/12 ]
Please have a look at this.
Comment by Mark Nunberg [ 05/Oct/12 ]
Michael,

I would not try this test manually.. the use case in more detail is as follows:

- Single CouchbaseClient object
- 20 user threads. 10 setting and 10 getting the same sorts of kv
- Operations are done asynchronously. They are submitted into a queue which is then checked periodically for isDone/isCancelled.
- 4 node cluster. Nodes are removed, connections are broken

The issue is those polling methods never returning true, unless they are retrieved synchronously (i.e. ft.get()).. which is actually an accidental detail
Comment by Matt Ingenthron [ 24/Oct/12 ]
We looked at this pretty closely today. The issue here is that the client as designed relies on the get() from the caller to trigger the timeout. An operation will, somewhat correctly, never transition to isDone() or isCancelled() unless someone cares to use it.

The scenario that was likely in play over the WAN here is that the request was in flight to the server while the config was in flight down to the client. It arrives at the server, but is never responded to. Since the get() is never called, it'll never time out and transition to the canceled state.

We recommend you change the test code to use the queue more like a queue and just get() each one. Iterating through the queue is a bit funny in the first place, but if using the get() on the Future objects, you'll still have asynchronous behavior and much of the time the get() will be returning since the data is already there.
Comment by Matt Ingenthron [ 24/Oct/12 ]
This behavior should be better documented, both in the javadoc and in the API reference.




Generated at Thu Jul 31 13:51:42 CDT 2014 using JIRA 5.2.4#845-sha1:c9f4cc41abe72fb236945343a1f485c2c844dac9.