Keith Turner
Keith Turner
In Accumulo 2.0 summarizes were introduced. Fluo could use these to count column types and notifications. These counts could be use to make compaction decisions.
Fluo's notifications are currently placed in a separate locality group. This means when scanning for notifications, user data is not read. Notifications can only be garbage collected with a full...
Scanning for notifications is a continuous activity that Fluo does. It may make sense to do one of the following. * Have a dedicated thread pool for notifcation scans with...
For get jars command add in #907 the command expects the target directory to exist but does not care if the directory is non-empty. I found myself wondering if the...
The two scenarios below really bother me. - User writes an observer that has conflicting dependencies with the Fluo worker thats running the observer. - User writes two observers that...
Saw the following error in travis when building #584 ``` Running io.fluo.integration.impl.CollisionIT Tests run: 1, Failures: 1, Errors: 0, Skipped: 0, Time elapsed: 45.833 sec
It would be useful if the fluo scan command had an option to produce machine readable output. This could be something like `fluo scan -a app1 --json` or `fluo scan...
Currently when an Observer fails an exception is logged and Fluo retries when it sees the notification again. What if anything can be done to improve this? Below are some...
Saw the following failure on TravisCI for #797 ``` Running org.apache.fluo.integration.impl.TransactorIT 2016-11-05 02:31:30,756 [imps.CuratorFrameworkImpl] ERROR: Background exception was not retry-able or retry gave up java.lang.IllegalStateException: Client is not started at...
The original intent of #626 was to support fetching multiple Spans. This was not done because the result may not fit into memory. At the time when #626 was worked...