View of currently running statements

[Aaron Molitor <am...@splicemachine.com>]

Hi all, I mn quite new to Trafodion and just trying to understand what information is available (and where it is) when I am running statements. 

I am running a simplified workflow based on the TPC-H benchmark.  The flow is to:

- create tables
- bulk load data
- count tables
- create indexes
- analyze/collect statistics
- run the 22 TPC-H queries

I am currently on the count tables step. The system got an rpc timeout exception at the client (trafci), and after dcscheck showed that the dcs master was down.

Two questions:
Should I increase the rpc timeout for hbase? If so, how far is reasonable?
How can I see what is running, and how it is progressing (sql statements in general, specifically count(*) and load statements)?

Here is the environment description:
9 nodes all are:
Dell R420
- 2 x E5-2430v2 2.5GHz (6C 12T)
- 64GB RAM
- 4x1TB SATA
Running 
- CDH 5.4.10 (ZooKeeper, HDFS, HBase, YARN Hive)
- Trafidion 2.0.1 (binary installer)

Thanks, 
Aaron

RE: View of currently running statements

[Selva Govindarajan <se...@esgyn.com>]

Thanks Aaron for trying out Trafodion.

You can refer the documentation at the Trafodion website
http://trafodion.apache.org/docs/sql_reference/index.html#sql_runtime_stat
istics and it has a whole section on how to obtain the runtime information
about the query once the query id is known.

You can find the query id of the currently running queries by issuing the
following commands after logging into the trafodion cluster

cd $MY_SQROOT/export/limited-support-tools/LSO
./offender -s active

You can also refer to README file at this location or issue ./offender
-help to get the information about many different usages of this feature.

Selva

-----Original Message-----
From: Aaron Molitor [mailto:amolitor@splicemachine.com]
Sent: Sunday, September 18, 2016 9:33 AM
To: user@trafodion.incubator.apache.org
Subject: View of currently running statements

Hi all, I mn quite new to Trafodion and just trying to understand what
information is available (and where it is) when I am running statements.

I am running a simplified workflow based on the TPC-H benchmark.  The flow
is to:

- create tables
- bulk load data
- count tables
- create indexes
- analyze/collect statistics
- run the 22 TPC-H queries

I am currently on the count tables step. The system got an rpc timeout
exception at the client (trafci), and after dcscheck showed that the dcs
master was down.

Two questions:
Should I increase the rpc timeout for hbase? If so, how far is reasonable?
How can I see what is running, and how it is progressing (sql statements
in general, specifically count(*) and load statements)?

Here is the environment description:
9 nodes all are:
Dell R420
- 2 x E5-2430v2 2.5GHz (6C 12T)
- 64GB RAM
- 4x1TB SATA
Running
- CDH 5.4.10 (ZooKeeper, HDFS, HBase, YARN Hive)
- Trafidion 2.0.1 (binary installer)

Thanks,
Aaron