Hi Ben,
The one thing that looks strange in the screenshot is the
ExecuteSqlCommand having FlowFiles queued in its incoming connection.
Those should be transferred to 'failure' relationship.
Following executeSql() method, shouldn't it re-throw the caught exception?
try (Connection
Hi Koji, no problem. You could check the code of processor WaitBatch at the
link:
https://drive.google.com/open?id=1DMpW5GMiXpyZQdui989Rr3D9rlchQfWQ
I also uploaded a snapshot of part of NiFi flow which includes the
ExecuteSqlCommand and WaitBatch, you could check the picture at the link:
Hi Ben,
I was referring these two log messages in your previous email.
These two messages are both written by ExecuteSqlCommand, it does not
mean 'it was executed again'.
```
2017-12-26 07:00:01,312 INFO [Timer-Driven Process Thread-1]
c.z.nifi.processors.ExecuteSqlCommand
Hi Koji, I will print the sql before actually executing it, but I checked
the error log line you mentioned in your reply, this error was thrown by
NiFi from within another processor called WaitBatch.
I didn't find similar errors as the one from the ExecuteSqlCommand
processor, I think it's because
Hi Ben,
The following two log messages are very close in terms of written
timestamp, but have different log level.
2017-12-26 07:00:01,312 INFO
2017-12-26 07:00:01,315 ERROR
I guess those are logged within a single onTrigger of your
ExecuteSqlCommand custom processor, one is before executing,
Thanks Koji, I will look into this article about the record model.
By the way, that error I previously mentioned to you occurred again, I
could see the sql query was executed twice in the log, this time I had
turned on the verbose NiFi logging, the sql query is as below:
2017-12-26 07:00:01,312
Hi Ben,
This blog post written by Mark, would be a good starting point to get
familiar with NiFi Record model.
https://blogs.apache.org/nifi/entry/record-oriented-data-with-nifi
HA for DistributedMapCacheClientService and DistributedMapCacheServer
pair is not supported at the moment. If you need
Steven
Understood. That it now shows you this is a good thing in 1.4.0
because what was happening in the older versions is one of the sets of
processors was being loaded. You are probably seeing a warning in the
app log at startup. So depending on which nar is loaded first
(standard or your
Yes, I agree. Unfortunately, our schedule may not allow for that at this time
so we will probably have to put off migrating for now.
Thanks to you and Bryan for your replies.
Thank you,
Steven K. Byers
DXC Technology - Contractor
Software Developer - Joint Medical Logistics Functional
Steven
To Bryan's point if it will be best if you simply copy the code/base
classes over to your Nar rather than extending from existing concrete
implemented processors.
Thanks
On Tue, Dec 26, 2017 at 1:50 PM, Byers, Steven K (Steve) CTR USARMY
MEDCOM JMLFDC (US)
Bryan,
We have extended a few processors from standard processors. For example, one
of the processors we are extending is:
org.apache.nifi.processors.standard.JmsConsumer
In another processor we are importing
org.apache.nifi.processors.standard.util.JdbcCommon.
I tried making the
Can you give some more details about what the dependency is for?
If it is some utility code that exists in standard processors then we
should be looking to move that to other reusable modules so that you can
depend on it without depending on the processors.
If it is because you extended a
Thanks for the reply, Bryan,
Yes, two of our custom processors have a dependency on the standard processors.
The dependency cannot be removed or those processors will not compile.
Thank you,
Steven K. Byers
DXC Technology - Contractor
Software Developer - Joint Medical Logistics Functional
Hello,
This means your custom NAR is bundling the standard processors jar and as a
result they are getting discovered twice, once from your NAR and once from
the standard NAR.
You’ll have to look at your maven dependencies for your custom NARs and
figure out why the dependency on standard
Hi devs,
I'm looking into moving from NiFi 1.1.2 to 1.4.0. We have several custom
processors and services. Everything is compiling without any problems but
when I put the services into the 1.4.0 instance, NiFi shows in the list of
processors a 1.1.2 and 1.4.0 version of all processors including
Thanks for your quick response, Koji, I haven't heard and seen anything
about the NiFi record data model when I was reading the NiFi
documentations,could you tell me where this model is documented? Thanks.
By the way, to my knowledge, when you need to use the DistributedMapCacheServer
from
Hi Ben,
As you found from existing code, DistributedMapCache is used to share
state among different processors, and it can be used by your custom
processors, too.
However, I'd recommend to avoid such tight dependencies between
FlowFiles if possible, or minimize the part in flow that requires that
Hi Milan,
Thanks for your contribution! I reviewed the PR and posted a comment there.
Would you check that?
Koji
On Sat, Dec 23, 2017 at 7:15 AM, Milan Das wrote:
> I have logged a defect in NIFI. ListS3 is generation duplicate flows when
> S3 throughput is high.
>
>
>
>
Hi guys, I'm currently trying to find a proper way in nifi which could sync
status between my custom processors.
our requirement is like this, we're doing some ETL work using nifi and I'm
extracting the data from DB into batches of FlowFiles(each batch of
FlowFile has a flag FlowFile indicating
19 matches
Mail list logo