[
https://issues.apache.org/jira/browse/KAFKA-16858?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17850375#comment-17850375
]
Adam Strickland commented on KAFKA-16858:
-----------------------------------------
{quote}Do you see this fault with another SMT in the chain ahead of Flatten?
{quote}
No. It consistently shows up only from the Flatten SMT and always on the same
attribute.
{quote}What Converter (sink connector) or source connector implementation are
you using?
{quote}
This is a sink connector, writing to Postgres using
`io.confluent.connect.jdbc.JdbcSinkConnector`. The value converter is
`io.confluent.connect.protobuf.ProtobufConverter`.
Some other data points:
* We use this same SMT in multiple connectors; we see the NPE in only one of
them.
* While it always throws from the same attribute, it doesn't *always* throw...
For instance, the NPE was logged 52 times within the last 24 hours, out of 234
messages processed via that connector.
* The connector we see it on has multiple SMTs; the Flatten is the 4th SMT
applied.
* For reference, the SMTs being applied prior to Flatten are, in order:
## `io.confluent.connect.transforms.Filter$Value`,
## `o.a.k.connect.transforms.ExtravtField$Value`, and
## another `o.a.k.connect.transforms.ExtravtField$Value`
* The attribute it breaks on is an array of `Struct`s; the array can be empty
and the attribute is itself optional. We have seen the break occur in 3 states
(attribute missing, attribute present with empty list, and attribute present
with a list). We have not seen it occur when the attribute is present but the
value is `null`. However, I think that's simply a use case we do not have.
* We have not used the trace logger you linked to, but at one point inserted
our own logging statements at various levels in our custom transform; this was
how we found the faulty attribute. One other thing we have noticed is the stack
recurses all the way down into `ConnectSchema.validateValue` an unexpected
number of times, resolving to `Schema\{STRUCT}` some number of times prior to
returning `null`. The number of times is not consistent and does not correlate
(as far as we can tell) with the number of items in the array or any other
measure we have checked.
> Flatten SMT throws NPE
> ----------------------
>
> Key: KAFKA-16858
> URL: https://issues.apache.org/jira/browse/KAFKA-16858
> Project: Kafka
> Issue Type: Bug
> Components: connect
> Affects Versions: 3.6.0
> Environment: Kafka 3.6 by way of CP 7.6.0
> Reporter: Adam Strickland
> Priority: Major
> Attachments: FlattenTest.java
>
>
> {{ConnectSchema.expectedClassesFor}} sometimes will throw an NPE as part of a
> call to an SMT chain. Stack trace snippet:
> {{at
> com.github.momenttechnology.kafka.connect.transforms.MomentFlatten.apply(MomentFlatten.java:84)}}
> {{at
> com.github.momenttechnology.kafka.connect.transforms.MomentFlatten.applyWithSchema(MomentFlatten.java:173)}}
> {{at
> com.github.momenttechnology.kafka.connect.transforms.MomentFlatten.buildWithSchema(MomentFlatten.java:280)}}
> {{at
> com.github.momenttechnology.kafka.connect.transforms.MomentFlatten.buildWithSchema(MomentFlatten.java:280)}}
> {{at
> com.github.momenttechnology.kafka.connect.transforms.MomentFlatten.buildWithSchema(MomentFlatten.java:280)}}
> {{at
> com.github.momenttechnology.kafka.connect.transforms.MomentFlatten.buildWithSchema(MomentFlatten.java:274)}}
> {{at org.apache.kafka.connect.data.Struct.put(Struct.java:203)}}
> {{at org.apache.kafka.connect.data.Struct.put(Struct.java:216)}}
> {{at
> org.apache.kafka.connect.data.ConnectSchema.validateValue(ConnectSchema.java:255)}}
> {{at
> org.apache.kafka.connect.data.ConnectSchema.validateValue(ConnectSchema.java:213)}}
> {{at
> org.apache.kafka.connect.data.ConnectSchema.validateValue(ConnectSchema.java:224)}}
> {{at
> org.apache.kafka.connect.data.ConnectSchema.expectedClassesFor(ConnectSchema.java:268)}}
> (the above transform is a sub-class of
> {{{}o.a.k.connect.transforms.Flatten{}}}; have confirmed that the error
> occurs regardless).
> The field being transformed is an array of structs. If the call to
> {{Schema#valueSchema()}} (o.a.k.connect.data.ConnectSchema.java:255) returns
> {{{}null{}}}, the subsequent call to {{Schema#name()}} at
> o.a.k.connect.data.ConnectSchema:268 throws an NPE.
> The strange thing that we have observed is that this doesn't always happen;
> *sometimes* the struct's schema is found and sometimes it is not. We have
> been unable to determine the root cause, but have constructed a test that
> replicates the problem as observed (see attachment).
> In our case we have worked around the issue with the aforementioned sub-class
> of {{{}Flatten{}}}, catching and logging the NPE on that specific use-case.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
