Does this mean phoenix will start each thread to scan one region?
Despite phoenix.query.threadPoolSize is 128 ?
No wonder it fails every time.
在 2016年06月23日 00:44, Samarth Jain 写道:
Please look at this tuning guide: https://phoenix.apache.org/tuning.html
You probably would want to adjust these client side properties to deal
with your workload: phoenix.query.threadPoolSize and
phoenix.query.queueSize.
On Wed, Jun 22, 2016 at 9:34 AM, 金砖 <[email protected]
<mailto:[email protected]>> wrote:
16 regionservers, 1500+ regions.
each regionserver holds 100 + region, region splits equally to
each rs.
I ran into this error just after start a phoenix shell.
but scanning other tables with 32 regions did not get this problem
在 2016年06月23日 00:13, Alicia Shu 写道:
How many region servers in your cluster? How many regions in
each server?
This is a client side configuration.
Alicia
On 6/22/16, 12:07 AM, "金砖" <[email protected]
<mailto:[email protected]>> wrote:
Thanks Alicia Shu.
Is this client side error or server side error ?
I did not say anything suspicious in server logs.
PS: there are 1500+ regions in table.
在 2016年06月22日 04:24, Alicia Shu 写道:
You have exhausted the threads. Set
“phoenix.query.threadPoolSize” to a
large number, say 3000 or 5000.
Alicia
On 6/21/16, 3:26 AM, "金砖" <[email protected]
<mailto:[email protected]>> wrote:
while select from phoenix shell.
I receive error like this:
Error: Task
org.apache.phoenix.job.JobManager$InstrumentedJobFutureTask@3af37506
rejected from
org.apache.phoenix.job.JobManager$1@91c4a3f[Running,
pool
size = 128, active threads = 128, queued tasks =
5000, completed tasks
=
90487] (state=08000,code=101)
org.apache.phoenix.exception.PhoenixIOException: Task
org.apache.phoenix.job.JobManager$InstrumentedJobFutureTask@3af37506
rejected from
org.apache.phoenix.job.JobManager$1@91c4a3f[Running,
pool
size = 128, active threads = 128, queued tasks =
5000, completed tasks
=
90487]
After restart cluster, this error still occurs.
Why would this happened? And how can I avoid this
error?
