Vuk Ercegovac has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/11608 )

Change subject: IMPALA-7597: wraps retries around CatalogMetaProvider
......................................................................


Patch Set 4:

(12 comments)

Agreed that spreading this retry pattern is not pretty. Some thoughts:
- is it an improvement when compared to the user seeing these exceptions?
- perhaps it indicates that there are too many service-level methods. I'm in 
favor of funneling these through a common path, but I'd rather do that 
restructure separately. an example of this is MetadataOp which fwict is used by 
the hs2 entry point.

Other suggestions welcome.

http://gerrit.cloudera.org:8080/#/c/11608/4/fe/src/main/java/org/apache/impala/service/Frontend.java
File fe/src/main/java/org/apache/impala/service/Frontend.java:

http://gerrit.cloudera.org:8080/#/c/11608/4/fe/src/main/java/org/apache/impala/service/Frontend.java@660
PS4, Line 660:    * Keeps track of retries when handling 
InconsistentMetadataFetchExceptions.
> Maybe add more context and documentation. In which cases is this used etc.
Done


http://gerrit.cloudera.org:8080/#/c/11608/4/fe/src/main/java/org/apache/impala/service/Frontend.java@666
PS4, Line 666:     private String msg_;
> final
Done


http://gerrit.cloudera.org:8080/#/c/11608/4/fe/src/main/java/org/apache/impala/service/Frontend.java@671
PS4, Line 671:
> ..of..
Done


http://gerrit.cloudera.org:8080/#/c/11608/4/fe/src/main/java/org/apache/impala/service/Frontend.java@680
PS4, Line 680:       LOG.warn("Retried {}: {} (retry #{} of {})", msg_, 
exception.getMessage(),
> Should any of this bubble up to the runtime profiles (if one exists) for th
would make sense to handle these uniformly with queries. added a todo. that 
plumbling is better done separately.


http://gerrit.cloudera.org:8080/#/c/11608/4/fe/src/main/java/org/apache/impala/service/Frontend.java@681
PS4, Line 681: );
> , exception) to log the trace and you can probably skip exception.getMessag
wasn't entirely sure what's the suggestion here, but I was aiming for 
consistency with the query-retry message


http://gerrit.cloudera.org:8080/#/c/11608/4/tests/custom_cluster/test_local_catalog.py
File tests/custom_cluster/test_local_catalog.py:

http://gerrit.cloudera.org:8080/#/c/11608/4/tests/custom_cluster/test_local_catalog.py@173
PS4, Line 173:   def test_cache_metrics(self, unique_database):
> did something change here or was it just moved to the top? Kind of confusin
separated these two types of tests into two classes so this test moved up here.


http://gerrit.cloudera.org:8080/#/c/11608/4/tests/custom_cluster/test_local_catalog.py@218
PS4, Line 218: _find_inconsistent_metadata
> name seems confusing. Maybe call check_metadata_fetch_retries or some such
renamed to _check_metadata_retries


http://gerrit.cloudera.org:8080/#/c/11608/4/tests/custom_cluster/test_local_catalog.py@221
PS4, Line 221: ,
> mention the randomness part?
Done


http://gerrit.cloudera.org:8080/#/c/11608/4/tests/custom_cluster/test_local_catalog.py@236
PS4, Line 236:       def stress_thread(client):
> Add a comment on what this does?
Done


http://gerrit.cloudera.org:8080/#/c/11608/4/tests/custom_cluster/test_local_catalog.py@250
PS4, Line 250:                 inconsistent_seen[0] += 1
> catch and record other exceptions since this runs in a thread?
Done


http://gerrit.cloudera.org:8080/#/c/11608/4/tests/custom_cluster/test_local_catalog.py@258
PS4, Line 258: 5
> I doubt if this is enough, I noticed some concurrency errors required > 60s
I saw it repro at this level, but I upped in case its too close to the edge. 
Set it to 30, which is the level used in tests already.


http://gerrit.cloudera.org:8080/#/c/11608/4/tests/custom_cluster/test_local_catalog.py@280
PS4, Line 280:       "alter table %s.%s set tblproperties ('numRows'='3')" %
> Run these on 'functional' tables with a concurrent refresh? That way you ca
I didn't want to modify the state (see the alter on L280). That can leak to 
other tests, which is a mess.



--
To view, visit http://gerrit.cloudera.org:8080/11608
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I43a21571d54a7966c5c68bea1fe69dbc62be2a0b
Gerrit-Change-Number: 11608
Gerrit-PatchSet: 4
Gerrit-Owner: Vuk Ercegovac <[email protected]>
Gerrit-Reviewer: Bharath Vissapragada <[email protected]>
Gerrit-Reviewer: Impala Public Jenkins <[email protected]>
Gerrit-Reviewer: Tianyi Wang <[email protected]>
Gerrit-Reviewer: Todd Lipcon <[email protected]>
Gerrit-Reviewer: Vuk Ercegovac <[email protected]>
Gerrit-Comment-Date: Wed, 17 Oct 2018 17:36:05 +0000
Gerrit-HasComments: Yes

Reply via email to