Hi guys, great job with the latest release. I do ...
# ingestion
f
Hi guys, great job with the latest release. I do have a different behavior with lineage between charts and datasets. I see the sources here, but not on the lineage graph.
m
@faint-hair-91313 is this something you are seeing in the new release and used to work in the old release? @green-football-43791
f
Yes.
Just upgraded and was trying different things. Including the STATS for Oracle.
By the way, I see you are doing sampling as in count rows < 1000, but Oracle has dictionaries that could possibly be queried instead of this. Would save a lot of time.
m
Sounds good. We’ll take a look. Did stats for oracle work?
g
@faint-hair-91313 does the lineage graph work for other entities?
m
Good point, we are using great expectations as “least common denominator” approach for profiling. We’ll probably need to add short cuts for specific sources that have inbuilt support.
f
Stats for Oracle. Yes, they've worked. See a snap here.
Lineage between datasets works ok.
g
Are you seeing lineage issues with all charts?
f
Had to look at 10 or so.. Yes. All have issues.
But the lineage is tracked in the main screen, not the graph.
g
Did you re-ingest chart data recently by any chance?
f
I did a clean install. Nuked everything and started over.
So yes 🙂
g
Ah ok- it seems like your chart relationships did not make it to the graph service
are you using neo4j or elastic?
f
Let me see ...
Copy code
bash-4.4$ sudo docker ps -a
CONTAINER ID   IMAGE                                       COMMAND                  CREATED       STATUS                      PORTS                                                                                      NAMES
a16edb15dda9   linkedin/datahub-kafka-setup:head           "/bin/sh -c ./kafka-…"   2 hours ago   Exited (0) 35 minutes ago                                                                                              kafka-setup
8d14304888b5   confluentinc/cp-schema-registry:5.4.0       "/etc/confluent/dock…"   2 hours ago   Up 36 minutes               0.0.0.0:8081->8081/tcp, :::8081->8081/tcp                                                  schema-registry
75c8ddea67b5   linkedin/datahub-frontend-react:head        "datahub-frontend/bi…"   2 hours ago   Up 36 minutes (healthy)     0.0.0.0:9002->9002/tcp, :::9002->9002/tcp                                                  datahub-frontend-react
c401305f99ae   confluentinc/cp-kafka:5.4.0                 "/etc/confluent/dock…"   2 hours ago   Up 36 minutes               0.0.0.0:9092->9092/tcp, :::9092->9092/tcp, 0.0.0.0:29092->29092/tcp, :::29092->29092/tcp   broker
9ea5bb0e43fd   linkedin/datahub-gms:head                   "/bin/sh -c /datahub…"   2 hours ago   Up 36 minutes (healthy)     0.0.0.0:8080->8080/tcp, :::8080->8080/tcp                                                  datahub-gms
ebe9a46d13a9   acryldata/datahub-mysql-setup:head          "dockerize /bin/sh -…"   2 hours ago   Exited (0) 36 minutes ago                                                                                              mysql-setup
61d1c1be9f01   linkedin/datahub-elasticsearch-setup:head   "dockerize /bin/sh -…"   2 hours ago   Exited (0) 35 minutes ago                                                                                              elasticsearch-setup
a90acb427bdd   confluentinc/cp-zookeeper:5.4.0             "/etc/confluent/dock…"   2 hours ago   Up 36 minutes               2888/tcp, 0.0.0.0:2181->2181/tcp, :::2181->2181/tcp, 3888/tcp                              zookeeper
8e266e2f3292   mysql:5.7                                   "docker-entrypoint.s…"   2 hours ago   Up 36 minutes               0.0.0.0:3306->3306/tcp, :::3306->3306/tcp, 33060/tcp                                       mysql
d0592890fbb2   elasticsearch:7.9.3                         "/tini -- /usr/local…"   2 hours ago   Up 36 minutes (healthy)     0.0.0.0:9200->9200/tcp, :::9200->9200/tcp, 9300/tcp                                        elasticsearch
I would say elastic?
g
yup, looks like there's no neo4j container
how many entities did you re-ingest?
f
All my charts, around 59.
g
and # of datasets?
trying to get a rough order of magnitude
f
3442
g
ok- lets try rebuilding your indexes
can you try running
Copy code
./docker/datahub-upgrade/datahub-upgrade.sh -u RestoreIndices
that should fix the lineage relationship
f
Ongoing ...
Didn't fix it.
g
hmm...
f
Copy code
...
Reading rows 16000 through 17000 from the aspects table.
Successfully sent MAEs for 17000 rows
Reading rows 17000 through 18000 from the aspects table.
Successfully sent MAEs for 18000 rows
Reading rows 18000 through 19000 from the aspects table.
Successfully sent MAEs for 19000 rows
Reading rows 19000 through 20000 from the aspects table.
Successfully sent MAEs for 19743 rows
Completed Step 4/4: SendMAEStep successfully.
Success! Completed upgrade with id RestoreIndices successfully.
Upgrade RestoreIndices completed with result SUCCEEDED. Exiting...
2021-08-04 14:26:27.814  INFO 1 --- [extShutdownHook] o.a.k.clients.producer.KafkaProducer     : [Producer clientId=producer-1] Closing the Kafka producer with timeoutMillis = 9223372036854775807 ms.
g
can you run
docker logs --tail datahub-gms
?
it might take a bit to process the re-index on gms side
if that is still printing out, let's give it some time
f
Copy code
sudo docker logs --tail datahub-gms
"docker logs" requires exactly 1 argument.
See 'docker logs --help'.

Usage:  docker logs [OPTIONS] CONTAINER

Fetch the logs of a container
Are you sure it's the right syntax.. Let me check ...
g
ah im sorry
--follow
(instead of --tail)
f
Log still running ... maybe it's still indexing?
g
sounds like it
f
Ok it stopped ...
Copy code
sbx2.tomo_sectors_hourly,PROD), urn:li:dataset:(urn:li:dataPlatform:oracle,edw.traffic_volume_h,PROD), urn:li:dataset:(urn:li:dataPlatform:oracle,edw.sco_sector_configuration_bb,PROD), urn:li:dataset:(urn:li:dataPlatform:oracle,edw.time_r,PROD)]
14:32:26.115 [pool-11-thread-1] INFO  c.l.metadata.filter.LoggingFilter - GET /entities?ids=List(urn%3Ali%3Adataset%3A%28urn%3Ali%3AdataPlatform%3Aoracle%2Cedw.sector_controlling_bb%2CPROD%29,urn%3Ali%3Adataset%3A%28urn%3Ali%3AdataPlatform%3Aoracle%2Csbx2.tomo_sectors_hourly%2CPROD%29,urn%3Ali%3Adataset%3A%28urn%3Ali%3AdataPlatform%3Aoracle%2Cedw.traffic_volume_h%2CPROD%29,urn%3Ali%3Adataset%3A%28urn%3Ali%3AdataPlatform%3Aoracle%2Cedw.sco_sector_configuration_bb%2CPROD%29,urn%3Ali%3Adataset%3A%28urn%3Ali%3AdataPlatform%3Aoracle%2Cedw.time_r%2CPROD%29) - batchGet - 200 - 10ms
14:32:26.121 [pool-11-thread-1] INFO  c.l.metadata.filter.LoggingFilter - GET /lineage?direction=INCOMING&urn=urn%3Ali%3Adataset%3A%28urn%3Ali%3AdataPlatform%3Aoracle%2Cedw.traffic_volume_h%2CPROD%29 - get - 200 - 3ms
14:32:26.122 [pool-11-thread-1] INFO  c.l.metadata.filter.LoggingFilter - GET /lineage?direction=OUTGOING&urn=urn%3Ali%3Adataset%3A%28urn%3Ali%3AdataPlatform%3Aoracle%2Cedw.traffic_volume_h%2CPROD%29 - get - 200 - 4ms
14:32:26.124 [pool-11-thread-1] INFO  c.l.metadata.filter.LoggingFilter - GET /lineage?direction=OUTGOING&urn=urn%3Ali%3Adataset%3A%28urn%3Ali%3AdataPlatform%3Aoracle%2Cedw.sector_controlling_bb%2CPROD%29 - get - 200 - 2ms
14:32:26.125 [pool-11-thread-1] INFO  c.l.metadata.filter.LoggingFilter - GET /lineage?direction=INCOMING&urn=urn%3Ali%3Adataset%3A%28urn%3Ali%3AdataPlatform%3Aoracle%2Cedw.sector_controlling_bb%2CPROD%29 - get - 200 - 3ms
14:32:26.127 [pool-11-thread-1] INFO  c.l.metadata.filter.LoggingFilter - GET /lineage?direction=OUTGOING&urn=urn%3Ali%3Adataset%3A%28urn%3Ali%3AdataPlatform%3Aoracle%2Cedw.time_r%2CPROD%29 - get - 200 - 2ms
14:32:26.129 [pool-11-thread-1] INFO  c.l.metadata.filter.LoggingFilter - GET /lineage?direction=INCOMING&urn=urn%3Ali%3Adataset%3A%28urn%3Ali%3AdataPlatform%3Aoracle%2Cedw.time_r%2CPROD%29 - get - 200 - 2ms
14:32:26.130 [pool-11-thread-1] INFO  c.l.metadata.filter.LoggingFilter - GET /lineage?direction=OUTGOING&urn=urn%3Ali%3Adataset%3A%28urn%3Ali%3AdataPlatform%3Aoracle%2Cedw.sco_sector_configuration_bb%2CPROD%29 - get - 200 - 2ms
14:32:26.133 [pool-11-thread-1] INFO  c.l.metadata.filter.LoggingFilter - GET /lineage?direction=INCOMING&urn=urn%3Ali%3Adataset%3A%28urn%3Ali%3AdataPlatform%3Aoracle%2Cedw.sco_sector_configuration_bb%2CPROD%29 - get - 200 - 3ms
14:32:26.133 [pool-11-thread-1] INFO  c.l.metadata.filter.LoggingFilter - GET /lineage?direction=OUTGOING&urn=urn%3Ali%3Adataset%3A%28urn%3Ali%3AdataPlatform%3Aoracle%2Csbx2.tomo_sectors_hourly%2CPROD%29 - get - 200 - 2ms
14:32:26.135 [pool-11-thread-1] INFO  c.l.metadata.filter.LoggingFilter - GET /lineage?direction=INCOMING&urn=urn%3Ali%3Adataset%3A%28urn%3Ali%3AdataPlatform%3Aoracle%2Csbx2.tomo_sectors_hourly%2CPROD%29 - get - 200 - 1ms
14:32:26.137 [qtp544724190-121] INFO  c.l.m.r.entity.EntityResource - BATCH GET [urn:li:dataset:(urn:li:dataPlatform:oracle,edw.sector_crossing_bbs,PROD), urn:li:dataset:(urn:li:dataPlatform:oracle,stg6.nm_volumes_sector,PROD), urn:li:dataset:(urn:li:dataPlatform:oracle,edw.sector_controlling_bbs,PROD), urn:li:dataset:(urn:li:dataPlatform:oracle,edw.sco_sector_configuration_bbs,PROD), urn:li:dataset:(urn:li:dataPlatform:oracle,inf.sector_configuration_f,PROD), urn:li:dataset:(urn:li:dataPlatform:oracle,inf.flight_sector_d,PROD)]
14:32:26.155 [pool-11-thread-1] INFO  c.l.metadata.filter.LoggingFilter - GET /entities?ids=List(urn%3Ali%3Adataset%3A%28urn%3Ali%3AdataPlatform%3Aoracle%2Cedw.sector_crossing_bbs%2CPROD%29,urn%3Ali%3Adataset%3A%28urn%3Ali%3AdataPlatform%3Aoracle%2Cstg6.nm_volumes_sector%2CPROD%29,urn%3Ali%3Adataset%3A%28urn%3Ali%3AdataPlatform%3Aoracle%2Cedw.sector_controlling_bbs%2CPROD%29,urn%3Ali%3Adataset%3A%28urn%3Ali%3AdataPlatform%3Aoracle%2Cedw.sco_sector_configuration_bbs%2CPROD%29,urn%3Ali%3Adataset%3A%28urn%3Ali%3AdataPlatform%3Aoracle%2Cinf.sector_configuration_f%2CPROD%29,urn%3Ali%3Adataset%3A%28urn%3Ali%3AdataPlatform%3Aoracle%2Cinf.flight_sector_d%2CPROD%29) - batchGet - 200 - 18ms
Maybe it was displaying on screen
g
those logs look like they were triggered by you interacting with the app
otherwise the log is silent?
f
Yes
b
Gratiel do you mind sending over your GMS debug log file?
f
I have an idea... we have to bypass a proxy, maybe neo4j couldn't be retrieved and then your script went for the previous installation. Give me a minute to try and get that one. Need to adapt your docker to use the proxy
Now, I see this little baby:
Copy code
b81e56beaa5b   mnexus001:8082/neo4j:4.0.6                    "/sbin/tini -g -- /d…"   49 seconds ago   Up 46 seconds                      0.0.0.0:7474->7474/tcp, :::7474->7474/tcp, 7473/tcp, 0.0.0.0:7687->7687/tcp, :::7687->7687/tcp   neo4j
Give 1 more min to run my ingestion.
Ah... damn, I am on an older version... d*** proxy again... I see gms:v0.8.6. Hold on, I'll get that fixed. Will tell you after.
g
ok- sounds good 👍
f
Copy code
Quickstarting DataHub: version v0.8.7
Datahub Neo4j volume found, starting with neo4j as graph service
g
ah ha.. that may explain it
f
Ok.. re-deployed everything with the new version, but still not graph relationship. Looking at the logs after running the restore indice I see this...
Copy code
15:17:24.391 [mae-consumer-job-client-0-C-1] ERROR c.l.m.k.MetadataAuditEventsProcessor - Error deserializing message: java.lang.NullPointerException
15:17:24.391 [mae-consumer-job-client-0-C-1] ERROR c.l.m.k.MetadataAuditEventsProcessor - Message: {"auditHeader": null, "oldSnapshot": null, "oldSystemMetadata": null, "newSnapshot": {"urn": "urn:li:dataset:(urn:li:dataPlatform:oracle,edw.flow_es,PROD)", "aspects": [{"platform": "urn:li:dataPlatform:oracle", "name": "edw.flow_es", "origin": "PROD"}]}, "newSystemMetadata": null, "operation": "UPDATE"}
15:17:24.391 [mae-consumer-job-client-0-C-1] INFO  c.l.m.k.MetadataAuditEventsProcessor - {urn=urn:li:dataset:(urn:li:dataPlatform:oracle,edw.flow_es,PROD), aspects=[{com.linkedin.metadata.key.DatasetKey={name=edw.flow_es, platform=urn:li:dataPlatform:oracle, origin=PROD}}, {com.linkedin.common.GlobalTags={tags=[{tag=urn:li:tag:Data Warehouse}]}}]}
15:17:24.399 [I/O dispatcher 1] INFO  c.l.m.s.e.update.BulkListener - Successfully fed bulk request. Number of events: 1 Took time ms: -1
15:17:24.400 [mae-consumer-job-client-0-C-1] ERROR c.l.m.k.MetadataAuditEventsProcessor - Error deserializing message: java.lang.NullPointerException
15:17:24.400 [mae-consumer-job-client-0-C-1] ERROR c.l.m.k.MetadataAuditEventsProcessor - Message: {"auditHeader": null, "oldSnapshot": null, "oldSystemMetadata": null, "newSnapshot": {"urn": "urn:li:dataset:(urn:li:dataPlatform:oracle,edw.flow_es,PROD)", "aspects": [{"platform": "urn:li:dataPlatform:oracle", "name": "edw.flow_es", "origin": "PROD"}, {"tags": [{"tag": "urn:li:tag:Data Warehouse"}]}]}, "newSystemMetadata": null, "operation": "UPDATE"}
Over and over...
g
Ok- taking a look.
b
ew!! this is no fun. we'll get back soon
f
Sure, thanks a lot!
g
got a fix here Gratiel- sorry about that!
f
Wow, so fast! Neat!
g
@faint-hair-91313 - we released the fix- could you upgrade and see if you can re-index now?
f
I am trying... fighting with the proxy ...
b
this dang proxy lol
f
Ok, sorted the proxy out... but still I do not see the relationships. For some reason it's still running on elastic.
g
Did you try running re-indexing after setting up again?
f
Yes.
And this time, no errors in the datahub-gms.
Regarding the proxy, I have to force getting the image from our local Nexus repository. If the Nexus repo does not find it, it will retrieve it from the internet. With the new version, you guys added an if case in the quickstart to handle the neo4j deployment. I missed that and that is why I couldn't get the new docker images. That's one. The pip library also comes via this Nexus repository. But there we had a failure timeout - everytime we try to dld a library that does not exist, you would have to try again in 24 hours before it could check. Several interactions with the IT team and got it sorted it out and reduced to 6 hours. 🤦‍♂️
g
Hey @faint-hair-91313 - update on this
Turns out there was a bug that was causing issues with Chart <> Dataset relationships
We've fixed this in the latest master 🙂
if you pull latest and re-index, you should be good to go!
Thank you for all your patience along with this bug.
f
No worries. You guys are doing am amazing job. Updated to 0.8.10 and relationships are back.
b
Thank you for looping back around Gratiel! You are an appreciated user. Keep the feedback coming!
g
Glad to hear!
🙌 1