curator-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jordan Zimmerman <jor...@jordanzimmerman.com>
Subject Re: Is this a known issue or fixed?
Date Tue, 29 Sep 2015 17:33:43 GMT
I think you mean Curator 2.7.0, right? This sounds similar to CURATOR-71 but I don’t know.
Please provide a test case and open an issue in Curator’s Jira.

-Jordan

On September 29, 2015 at 3:08:56 AM, Lu, Boying (boying.lu@emc.com) wrote:

Hi, All,  

We are using ZooKeepr 2.7.0 in our product and we found an issue in our daily testing.  
This issue is related to the InterProcessMutex.acquire() and I found following stack trace
in our log:  

org.apache.zookeeper.KeeperException$NoNodeException: KeeperErrorCode = NoNode for Sequential
path not found: _c_33115fca-5414-40bd-8dfe-baf3ab6f3c5a-lock-0000308430  
at org.apache.curator.framework.recipes.locks.StandardLockInternalsDriver.validateOurIndex(StandardLockInternalsDriver.java:81)
 
at org.apache.curator.framework.recipes.locks.StandardLockInternalsDriver.getsTheLock(StandardLockInternalsDriver.java:36)
 
at org.apache.curator.framework.recipes.locks.LockInternals.internalLockLoop(LockInternals.java:278)
 
at org.apache.curator.framework.recipes.locks.LockInternals.attemptLock(LockInternals.java:217)
 
at org.apache.curator.framework.recipes.locks.InterProcessMutex.internalLock(InterProcessMutex.java:232)
 
at org.apache.curator.framework.recipes.locks.InterProcessMutex.acquire(InterProcessMutex.java:108)
 

Reading some codes, I think the root cause maybe related to following snipped (in LockInternals.java:214):
 
try  
{  
ourPath = driver.createsTheLock(client, path, localLockNodeBytes);  
hasTheLock = internalLockLoop(startMillis, millisToWait, ourPath);  
}  
My understanding is that the "ourPath" is the full path of the created ephemeral node with
a sequential number,  
But the LockInternals.getSortedChildren() (referenced in the internalLockLoop()) doesn't has
any child with such sequential number,  
so the valid checking fails.  

Is this an known issue and is there any JIRA to track it?  

Thanks  

Boying  



Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message