[jira] [Commented] (IMPALA-7283) Consider backing off before retrying RPC

2018-07-11 Thread Michael Ho (JIRA)


[ 
https://issues.apache.org/jira/browse/IMPALA-7283?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16540930#comment-16540930
 ] 

Michael Ho commented on IMPALA-7283:


Not in the KRPC code but we can easily achieve that by rescheduling the RPC. We 
already do that for KrpcDataStreamSender.

 

 

> Consider backing off before retrying RPC
> 
>
> Key: IMPALA-7283
> URL: https://issues.apache.org/jira/browse/IMPALA-7283
> Project: IMPALA
>  Issue Type: Improvement
>  Components: Distributed Exec
>Reporter: Tim Armstrong
>Priority: Major
>
> [~stiga-huang] suggested here 
> https://gerrit.cloudera.org/#/c/10744/11/be/src/service/client-request-state.cc@628
> {quote}
> What about sleep several seconds before the next retry like this?
>  for (int i = 0; i < 3; ++i, sleep(3))
> Usually it will increase success rate if there're network issues or the 
> target server is stuck temporarily.
> {quote}
> This seems worth considering but I don't have the knowledge to really 
> evaluate it
> cc [~kwho] [~sailesh]



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

-
To unsubscribe, e-mail: issues-all-unsubscr...@impala.apache.org
For additional commands, e-mail: issues-all-h...@impala.apache.org



[jira] [Commented] (IMPALA-7283) Consider backing off before retrying RPC

2018-07-11 Thread Tim Armstrong (JIRA)


[ 
https://issues.apache.org/jira/browse/IMPALA-7283?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16540920#comment-16540920
 ] 

Tim Armstrong commented on IMPALA-7283:
---

Does KRPC have some kind of better retry logic?


> Consider backing off before retrying RPC
> 
>
> Key: IMPALA-7283
> URL: https://issues.apache.org/jira/browse/IMPALA-7283
> Project: IMPALA
>  Issue Type: Improvement
>  Components: Distributed Exec
>Reporter: Tim Armstrong
>Priority: Major
>
> [~stiga-huang] suggested here 
> https://gerrit.cloudera.org/#/c/10744/11/be/src/service/client-request-state.cc@628
> {quote}
> What about sleep several seconds before the next retry like this?
>  for (int i = 0; i < 3; ++i, sleep(3))
> Usually it will increase success rate if there're network issues or the 
> target server is stuck temporarily.
> {quote}
> This seems worth considering but I don't have the knowledge to really 
> evaluate it
> cc [~kwho] [~sailesh]



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

-
To unsubscribe, e-mail: issues-all-unsubscr...@impala.apache.org
For additional commands, e-mail: issues-all-h...@impala.apache.org



[jira] [Commented] (IMPALA-7283) Consider backing off before retrying RPC

2018-07-11 Thread Michael Ho (JIRA)


[ 
https://issues.apache.org/jira/browse/IMPALA-7283?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16540913#comment-16540913
 ] 

Michael Ho commented on IMPALA-7283:


Yes, the sleep is already there for {{QueryState::ReportExecStatusAux()}} 
{noformat}
  for (int i = 0; i < 3; ++i) {
ImpalaBackendConnection 
client(ExecEnv::GetInstance()->impalad_client_cache(),
query_ctx().coord_address, _status);
if (client_status.ok()) {
  rpc_status = client.DoRpc(::ReportExecStatus, params, 
);
  if (rpc_status.ok()) break;
}
if (i < 2) SleepForMs(FLAGS_report_status_retry_interval_ms);
  }
{noformat}

We can consider moving it into {{client-request-state.cc}} but most of those 
RPCs will be ported to KRPC in the near future so probably not worth doing more 
surgery on the logic in that file.

> Consider backing off before retrying RPC
> 
>
> Key: IMPALA-7283
> URL: https://issues.apache.org/jira/browse/IMPALA-7283
> Project: IMPALA
>  Issue Type: Improvement
>  Components: Distributed Exec
>Reporter: Tim Armstrong
>Priority: Major
>
> [~stiga-huang] suggested here 
> https://gerrit.cloudera.org/#/c/10744/11/be/src/service/client-request-state.cc@628
> {quote}
> What about sleep several seconds before the next retry like this?
>  for (int i = 0; i < 3; ++i, sleep(3))
> Usually it will increase success rate if there're network issues or the 
> target server is stuck temporarily.
> {quote}
> This seems worth considering but I don't have the knowledge to really 
> evaluate it
> cc [~kwho] [~sailesh]



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

-
To unsubscribe, e-mail: issues-all-unsubscr...@impala.apache.org
For additional commands, e-mail: issues-all-h...@impala.apache.org