[jira] Commented: (ZOOKEEPER-550) Java Queue Recipe

[Henry Robinson (JIRA)]

[ 
https://issues.apache.org/jira/browse/ZOOKEEPER-550?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=12764502#action_12764502
 ] 

Henry Robinson commented on ZOOKEEPER-550:
--

Looks good!

In offer(...) - would it be better to catch the NoNodeException and create the 
root then, rather than requiring an RPC every time?



 Java Queue Recipe
 -

 Key: ZOOKEEPER-550
 URL: https://issues.apache.org/jira/browse/ZOOKEEPER-550
 Project: Zookeeper
  Issue Type: New Feature
  Components: java client
Affects Versions: 3.2.1
Reporter: Steven Cheng
Priority: Minor
 Fix For: 3.3.0

 Attachments: ZOOKEEPER-550.patch


 This patch adds a recipe for creating a distributed queue with ZooKeeper 
 similar to the WriteLock recipe and some sequential tests.  This early 
 attempt follows the Java BlockingQueue interface, though it doesn't implement 
 it since I don't think there's a good reason for it to be Iterable.  

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.

[jira] Commented: (ZOOKEEPER-462) Last hint for open ledger

[Benjamin Reed (JIRA)]

[ 
https://issues.apache.org/jira/browse/ZOOKEEPER-462?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=12764535#action_12764535
 ] 

Benjamin Reed commented on ZOOKEEPER-462:
-

if the client asks all the bookies he may not be able to get the last committed 
entry if we allow for failures. the safest thing to do would be to get the last 
entry from each bookie and then use the entry id in the last committed field. 
that would mean that you would never be able to see the actual last committed 
record.

i think it would be good to allow the client to specify the last committed 
entry on the open. that way we allow the client to figure out the last 
committed record any way it wants, via communication from other processes for 
example, and it would keep the open code simple: it would just use the id it 
wouldn't need to worry about recovery.

 Last hint for open ledger
 -

 Key: ZOOKEEPER-462
 URL: https://issues.apache.org/jira/browse/ZOOKEEPER-462
 Project: Zookeeper
  Issue Type: New Feature
  Components: contrib-bookkeeper
Reporter: Flavio Paiva Junqueira
Assignee: Flavio Paiva Junqueira
 Fix For: 3.3.0

 Attachments: ZOOKEEPER-462.patch


 In some use cases of BookKeeper, it is useful to be able to read from a 
 ledger before closing the ledger. To enable such a feature, the writer has to 
 be able to communicate to a reader how many entries it has been able to write 
 successfully. The main idea of this jira is to continuously update a znode 
 with the number of successful writes, and a reader can, for example, watch 
 the node for changes.
  I was thinking of having a configuration parameter to state how often a 
 writer should update the hint on ZooKeeper (e.g., every 1000 requests, every 
 10,000 requests). Clearly updating more often increases the overhead of 
 writing to ZooKeeper, although the impact on the performance of writes to 
 BookKeeper should be minimal given that we make an asynchronous call to 
 update the hint.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.