If it's a Hector thing you may have better luck on the Hector user group. 

http://groups.google.com/group/hector-users

Cheers

-----------------
Aaron Morton
Freelance Developer
@aaronmorton

On 10/03/2012, at 8:33 AM, Daning Wang wrote:

Thanks Maciej. we have default value for retryDownedHostsDelayInSeconds. I think it is not about how long it checks the downed host, I suspect the HostRetryService is down. Below is the very first exception, what does this message mean  - " HConnectionManager returned a null client after aquisition - are we shutting down?"



2012-03-08 16:37:15,103 [pool-2-thread-34288]   Cassandra client acquisition interrupted
java.lang.InterruptedException
        at java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.reportInterruptAfterWait(Unknown Source)
        at java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.awaitNanos(Unknown Source)
        at java.util.concurrent.ArrayBlockingQueue.poll(Unknown Source)
        at me.prettyprint.cassandra.connection.ConcurrentHClientPool.waitForConnection(ConcurrentHClientPool.java:117)
        at me.prettyprint.cassandra.connection.ConcurrentHClientPool.borrowClient(ConcurrentHClientPool.java:77)
        at me.prettyprint.cassandra.connection.HConnectionManager.operateWithFailover(HConnectionManager.java:226)
        at me.prettyprint.cassandra.model.ExecutingKeyspace.doExecuteOperation(ExecutingKeyspace.java:97)
        at me.prettyprint.cassandra.model.CqlQuery.execute(CqlQuery.java:93)
        at com.netseer.cassandra.cache.dao.CacheReader.getRows(CacheReader.java:267)
        at com.netseer.cassandra.cache.dao.CacheReader.getCache0(CacheReader.java:55)
        at com.netseer.cassandra.cache.dao.CacheDao.getCaches(CacheDao.java:85)
        at com.netseer.cassandra.cache.dao.CacheDao.getCache(CacheDao.java:71)
        at com.netseer.cassandra.cache.dao.CacheDao.getCache(CacheDao.java:149)
        at com.netseer.cassandra.cache.service.CacheServiceImpl.getCache(CacheServiceImpl.java:55)
        at com.netseer.cassandra.cache.service.CacheServiceImpl.getCache(CacheServiceImpl.java:28)
        at com.netseer.dsat.cache.CassandraDSATCacheImpl.get(CassandraDSATCacheImpl.java:62)
        at com.netseer.dsat.cache.CassandraDSATCacheImpl.getTimedValue(CassandraDSATCacheImpl.java:144)
        at com.netseer.dsat.serving.GenericCacheManager$4.call(GenericCacheManager.java:427)
        at com.netseer.dsat.serving.GenericCacheManager$4.call(GenericCacheManager.java:1)
        at java.util.concurrent.FutureTask$Sync.innerRun(Unknown Source)
        at java.util.concurrent.FutureTask.run(Unknown Source)
        at java.util.concurrent.ThreadPoolExecutor$Worker.runTask(Unknown Source)
        at java.util.concurrent.ThreadPoolExecutor$Worker.run(Unknown Source)
        at java.lang.Thread.run(Unknown Source)
2012-03-08 16:37:15,104 [pool-2-thread-34288]   Failed getting remote cache for key=Key String = 'http://www.my-banners.com', long key = 5630311119483252185, keyType = 'PATH'
me.prettyprint.hector.api.exceptions.HectorException: HConnectionManager returned a null client after aquisition - are we shutting down?
        at me.prettyprint.cassandra.connection.ConcurrentHClientPool.borrowClient(ConcurrentHClientPool.java:83)
        at me.prettyprint.cassandra.connection.HConnectionManager.operateWithFailover(HConnectionManager.java:226)
        at me.prettyprint.cassandra.model.ExecutingKeyspace.doExecuteOperation(ExecutingKeyspace.java:97)
        at me.prettyprint.cassandra.model.CqlQuery.execute(CqlQuery.java:93)



On Mon, Mar 5, 2012 at 10:56 PM, Maciej Miklas <mac.miklas@googlemail.com> wrote:
Have you tried to change: me.prettyprint.cassandra.service.CassandraHostConfigurator#retryDownedHostsDelayInSeconds ?

Hector will ping down hosts every xx seconds and recover connection.

Regards,
Maciej


On Mon, Mar 5, 2012 at 8:13 PM, Daning Wang <daning@netseer.com> wrote:
I just got this error ": All host pools marked down. Retry burden pushed out to client." in a few clients recently, client could not  recover, we have to restart client application.  we are using 0.8.0.3 hector.

At that time we did compaction  for a CF, it takes several hours, server was busy. But I think client should recover after server load was down.

Any bug reported about this? I did search but could not find one.

Thanks,

Daning