[
https://issues.apache.org/jira/browse/PHOENIX-6273?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Saksham Gangwar updated PHOENIX-6273:
-------------------------------------
Description:
Recently we switched an MR application from scanning live tables to scanning
snapshots (PHOENIX-3744). We ran into a severe performance issue, which turned
out to a correctness issue due to over-lapping scan splits generation. After
some debugging we figured that it has been fixed via PHOENIX-4997.
We also *need not restore the snapshot per map task*. Currently, we restore the
snapshot once per map task into a temp directory. For large tables on big
clusters, this creates a storm of NN RPCs. We can do this once per job and let
all the map tasks operate on the same restored snapshot. HBase already did this
via HBASE-18806, we can do something similar.
All other performance suggestions here:
https://issues.apache.org/jira/browse/PHOENIX-6081
was:
Recently we switched an MR application from scanning live tables to scanning
snapshots (PHOENIX-3744). We ran into a severe performance issue, which turned
out to a correctness issue due to over-lapping scan splits generation. After
some debugging we figured that it has been fixed via PHOENIX-4997.
We also *need not restore the snapshot per map task*. Currently, we restore the
snapshot once per map task into a temp directory. For large tables on big
clusters, this creates a storm of NN RPCs. We can do this once per job and let
all the map tasks operate on the same restored snapshot. HBase already did this
via HBASE-18806, we can do something similar.
> All the map tasks should operate on the same restored snapshot
> --------------------------------------------------------------
>
> Key: PHOENIX-6273
> URL: https://issues.apache.org/jira/browse/PHOENIX-6273
> Project: Phoenix
> Issue Type: Bug
> Components: core
> Affects Versions: 5.0.0, 4.14.3
> Reporter: Saksham Gangwar
> Assignee: Saksham Gangwar
> Priority: Major
> Fix For: 4.x, 4.16.1
>
>
> Recently we switched an MR application from scanning live tables to scanning
> snapshots (PHOENIX-3744). We ran into a severe performance issue, which
> turned out to a correctness issue due to over-lapping scan splits generation.
> After some debugging we figured that it has been fixed via PHOENIX-4997.
> We also *need not restore the snapshot per map task*. Currently, we restore
> the snapshot once per map task into a temp directory. For large tables on big
> clusters, this creates a storm of NN RPCs. We can do this once per job and
> let all the map tasks operate on the same restored snapshot. HBase already
> did this via HBASE-18806, we can do something similar.
>
> All other performance suggestions here:
> https://issues.apache.org/jira/browse/PHOENIX-6081
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
