I finally got to the bottom of things. There were two issues at play in my
particular environment.
1. An Ambari bug [0] means my spark-defaults.conf file was garbage. I
hardly thought of it when I hit the issue with MR job submission; its
impact on Spark was much more subtle.
2. YARN client
Hi guys,
We recently migrated one of our user facing use cases to HBase and we are using
Phoenix as query layer.
We managed to get a singles record in 30Ms.
Here is the response times breakdown.
75th - 29MS
95th - 43MS
99th - 76 MS
We have about 6Billion records in store and each row contains
That's great to hear. Looking forward to the doc patch!
On Wed, Jan 20, 2016 at 3:43 PM, Nick Dimiduk wrote:
> Josh -- I deployed my updated phoenix build across the cluster, added the
> phoenix-client-spark.jar to configs on the whole cluster, and now basic
> dataframe
Well, I spoke too soon. It's working, but in local mode only. When I invoke
`pyspark --master yarn` (or yarn-client), the submitted application goes
from ACCEPTED to FAILED, with a NumberFormatException [0] in my container
log. Now that Phoenix is on my classpath, I'm suspicious that the versions
Hi James,
Thanks for being willing to assist.
This is what the input data record will look like (test data) :
UserID
DateTime
TXNID
DeviceID
IPAddress
UsageArray
URIArray
12345678901
20151006124945
992194978
123456789012345
111.111.111.111
I have a similar data pattern and 100ms response time is fairly consistent.
I’ve been trying hard to find the right set of configs to get closer to 10-20ms
with no luck, but I’m finding that 100ms average is pretty reasonable.
From: Willem Conradie [mailto:willem.conra...@pbtgroup.co.za]
Sent: