[jira] [Updated] (FLINK-34491) Move from experimental support to production support for Java 17

2024-02-29 Thread Dhruv Patel (Jira)


 [ 
https://issues.apache.org/jira/browse/FLINK-34491?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Dhruv Patel updated FLINK-34491:

Description: 
This task is to move away from experimental support for Java 17 to production 
support so that teams running Flink in production can migrate to Java 17 
successfully

*Background:*

Flink supports protobuf dataformat to exchange messages between different 
operators and the serialization and deserialization of these protobufs messages 
are performed by library called "Kryo". In order to move away from experimental 
support of Java 17 released as part of Flink 1.18.1, the Kryo library in Flink 
1.18.1 needs to be updated from 2.24.0 to 5.5.0 because Kryo 2.24.0 does not 
support Java 17. This improvement plan is tracked as part of this ticket 
[https://cwiki.apache.org/confluence/display/FLINK/FLIP-317%3A+Upgrade+Kryo+from+2.24.0+to+5.5.0]

All Flink applications using protobuf currently generate state with Kryo v2. 
Once the above improvement plan is complete all Flink applications will fully 
support reading that state and write newer state with Kryo v5. However, latest 
Kryo v5 doesn't support snapshots made by a previous Kryo v2. This will prevent 
applications which are using snapshot mechanism to deploy their jobs to latest 
Flink version with Kryo v5 support without a bridge version running on Java 11. 
Applications will have to run on a bridge release version that will read their 
state with Kryo v2 data and write it with Kryo v5 data before upgrading to a 
future version of Flink that completely drops support for Kryo v2.

Basically, Flink applications using protobuf dataformat cannot move directly 
from Java 8 to Java 17 without downtime after the kryo v5 release in Flink. 
Applications will need to first move to Java 11 (bridging version) and then 
move to Java 17 to have a safe deployment.

*Blocker for this task:*

Upgrade to Kryo 5.5.0 which supports Java 17 and a path for snapshot migration
https://issues.apache.org/jira/browse/FLINK-3154.

  was:
This task is to move away from experimental support for Java 17 to production 
support so that teams running Flink in production can migrate to Java 17 
successfully

*Background:*

Flink supports protobuf dataformat to exchange messages between different 
operators and the serialization and deserialization of those protobufs are 
performed by library called "Kryo". In order to move away from experimental 
support of Java 17 released as part of Flink 1.18.1, the Kryo library in Flink 
1.18.1 needs to be updated from 2.24.0 to 5.5.0 because Kryo 2.24.0 does not 
support Java 17. This improvement plan is tracked as part of this ticket 
[https://cwiki.apache.org/confluence/display/FLINK/FLIP-317%3A+Upgrade+Kryo+from+2.24.0+to+5.5.0]

All Flink applications using protobuf currently generate state with Kryo v2. 
Once the above improvement plan is complete all Flink applications will fully 
support reading that state and write newer state with Kryo v5. However, latest 
Kryo v5 doesn't support snapshots made by a previous Kryo v2. This will prevent 
applications which are using snapshot mechanism to deploy their jobs to latest 
Flink version with Kryo v5 support without a bridge version running on Java 11. 
Applications will have to run on a bridge release version that will read their 
state with Kryo v2 data and write it with Kryo v5 data before upgrading to a 
future version of Flink that completely drops support for Kryo v2.

Basically, Flink applications using protobuf dataformat cannot move directly 
from Java 8 to Java 17 without downtime after the kryo v5 release in Flink. 
Applications will need to first move to Java 11 (bridging version) and then 
move to Java 17 to have a safe deployment.

*Blocker for this task:*

Upgrade to Kryo 5.5.0 which supports Java 17 and a path for snapshot migration
https://issues.apache.org/jira/browse/FLINK-3154.


> Move from experimental support to production support for Java 17
> 
>
> Key: FLINK-34491
> URL: https://issues.apache.org/jira/browse/FLINK-34491
> Project: Flink
>  Issue Type: New Feature
>Affects Versions: 1.18.1
>Reporter: Dhruv Patel
>Priority: Major
>
> This task is to move away from experimental support for Java 17 to production 
> support so that teams running Flink in production can migrate to Java 17 
> successfully
> *Background:*
> Flink supports protobuf dataformat to exchange messages between different 
> operators and the serialization and deserialization of these protobufs 
> messages are performed by library called "Kryo". In order to move away from 
> experimental support of Java 17 released as part of Flink 1.18.1, the Kryo 
> library in Flink 1.18.1 needs to be updated from 2.24.0 to 5.5.0 because Kryo 
> 2.24.0 does not support Java 

[jira] [Updated] (FLINK-34491) Move from experimental support to production support for Java 17

2024-02-28 Thread Dhruv Patel (Jira)


 [ 
https://issues.apache.org/jira/browse/FLINK-34491?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Dhruv Patel updated FLINK-34491:

Description: 
This task is to move away from experimental support for Java 17 to production 
support so that teams running Flink in production can migrate to Java 17 
successfully

*Background:*

Flink supports protobuf dataformat to exchange messages between different 
operators and the serialization and deserialization of those protobufs are 
performed by library called "Kryo". In order to move away from experimental 
support of Java 17 released as part of Flink 1.18.1, the Kryo library in Flink 
1.18.1 needs to be updated from 2.24.0 to 5.5.0 because Kryo 2.24.0 does not 
support Java 17. This improvement plan is tracked as part of this ticket 
[https://cwiki.apache.org/confluence/display/FLINK/FLIP-317%3A+Upgrade+Kryo+from+2.24.0+to+5.5.0]

All Flink applications using protobuf currently generate state with Kryo v2. 
Once the above improvement plan is complete all Flink applications will fully 
support reading that state and write newer state with Kryo v5. However, latest 
Kryo v5 doesn't support snapshots made by a previous Kryo v2. This will prevent 
applications which are using snapshot mechanism to deploy their jobs to latest 
Flink version with Kryo v5 support without a bridge version running on Java 11. 
Applications will have to run on a bridge release version that will read their 
state with Kryo v2 data and write it with Kryo v5 data before upgrading to a 
future version of Flink that completely drops support for Kryo v2.

Basically, Flink applications using protobuf dataformat cannot move directly 
from Java 8 to Java 17 without downtime after the kryo v5 release in Flink. 
Applications will need to first move to Java 11 (bridging version) and then 
move to Java 17 to have a safe deployment.

*Blocker for this task:*

Upgrade to Kryo 5.5.0 which supports Java 17 and a path for snapshot migration
https://issues.apache.org/jira/browse/FLINK-3154.

  was:
This task is to move away from experimental support for Java 17 to production 
support so that teams running Flink in production can migrate to Java 17 
successfully

*Background:*

Flink supports protobuf dataformat to exchange messages between different 
operators and the serialization and deserialization of those protobufs are 
performed by library called "Kryo". In order to move away from experimental 
support of Java 17 released as part of Flink 1.18.1, the Kryo library in Flink 
1.18.1 needs to be updated from 2.24.0 to 5.5.0 because Kryo 2.24.0 does not 
support Java 17. This improvement plan is tracked as part of this ticket 
https://issues.apache.org/jira/browse/FLINK-3154.

All Flink applications using protobuf currently generate state with Kryo v2. 
Once the above improvement plan is complete all Flink applications will fully 
support reading that state and write newer state with Kryo v5. However, latest 
Kryo v5 doesn't support snapshots made by a previous Kryo v2. This will prevent 
applications which are using snapshot mechanism to deploy their jobs to latest 
Flink version with Kryo v5 support without a bridge version running on Java 11. 
Applications will have to run on a bridge release version that will read their 
state with Kryo v2 data and write it with Kryo v5 data before upgrading to a 
future version of Flink that completely drops support for Kryo v2.

Basically, Flink applications using protobuf dataformat cannot move directly 
from Java 8 to Java 17 without downtime after the kryo v5 release in Flink. 
Applications will need to first move to Java 11 (bridging version) and then 
move to Java 17 to have a safe deployment.

Migration Plan is documented here: 
[https://cwiki.apache.org/confluence/display/FLINK/FLIP-317%3A+Upgrade+Kryo+from+2.24.0+to+5.5.0]

*Blocker for this task:*

Upgrade to Kryo 5.5.0 which supports Java 17 and a path for snapshot migration
https://issues.apache.org/jira/browse/FLINK-3154.


> Move from experimental support to production support for Java 17
> 
>
> Key: FLINK-34491
> URL: https://issues.apache.org/jira/browse/FLINK-34491
> Project: Flink
>  Issue Type: New Feature
>Affects Versions: 1.18.1
>Reporter: Dhruv Patel
>Priority: Major
>
> This task is to move away from experimental support for Java 17 to production 
> support so that teams running Flink in production can migrate to Java 17 
> successfully
> *Background:*
> Flink supports protobuf dataformat to exchange messages between different 
> operators and the serialization and deserialization of those protobufs are 
> performed by library called "Kryo". In order to move away from experimental 
> support of Java 17 released as part of Flink 1.18.1, the Kryo library in 
> Flink 1.18.1 needs to be 

[jira] [Updated] (FLINK-34491) Move from experimental support to production support for Java 17

2024-02-28 Thread Dhruv Patel (Jira)


 [ 
https://issues.apache.org/jira/browse/FLINK-34491?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Dhruv Patel updated FLINK-34491:

Description: 
This task is to move away from experimental support for Java 17 to production 
support so that teams running Flink in production can migrate to Java 17 
successfully

*Background:*

Flink supports protobuf dataformat to exchange messages between different 
operators and the serialization and deserialization of those protobufs are 
performed by library called "Kryo". In order to move away from experimental 
support of Java 17 released as part of Flink 1.18.1, the Kryo library in Flink 
1.18.1 needs to be updated from 2.24.0 to 5.5.0 because Kryo 2.24.0 does not 
support Java 17. This improvement plan is tracked as part of this ticket 
https://issues.apache.org/jira/browse/FLINK-3154.

All Flink applications using protobuf currently generate state with Kryo v2. 
Once the above improvement plan is complete all Flink applications will fully 
support reading that state and write newer state with Kryo v5. However, latest 
Kryo v5 doesn't support snapshots made by a previous Kryo v2. This will prevent 
applications which are using snapshot mechanism to deploy their jobs to latest 
Flink version with Kryo v5 support without a bridge version running on Java 11. 
Applications will have to run on a bridge release version that will read their 
state with Kryo v2 data and write it with Kryo v5 data before upgrading to a 
future version of Flink that completely drops support for Kryo v2.

Basically, Flink applications using protobuf dataformat cannot move directly 
from Java 8 to Java 17 without downtime after the kryo v5 release in Flink. 
Applications will need to first move to Java 11 (bridging version) and then 
move to Java 17 to have a safe deployment.

Migration Plan is documented here: 
[https://cwiki.apache.org/confluence/display/FLINK/FLIP-317%3A+Upgrade+Kryo+from+2.24.0+to+5.5.0]

*Blocker for this task:*

Upgrade to Kryo 5.5.0 which supports Java 17 and a path for snapshot migration
https://issues.apache.org/jira/browse/FLINK-3154.

  was:
This task is to move away from experimental support for Java 17 to production 
support so that teams running Flink in production can migrate to Java 17 
successfully

*Background:*

Flink supports protobuf dataformat to exchange messages between different 
operators and the serialization and deserialization of those protobufs are 
performed by library called "Kryo". In order to move away from experimental 
support of Java 17 released as part of Flink 1.18.1, the Kryo library in Flink 
1.18.1 needs to be updated from 2.24.0 to 5.5.0 because Kryo 2.24.0 does not 
support Java 17. This improvement plan is tracked as part of this ticket 
https://issues.apache.org/jira/browse/FLINK-3154.

All Flink applications using protobuf currently generate state with Kryo v2. 
Once the above improvement plan is complete all Flink applications will fully 
support reading that state and write newer state with Kryo v5. However, latest 
Kryo v5 doesn't support snapshots made by a previous Kryo v2. This will prevent 
applications which are using snapshot mechanism to deploy their jobs to latest 
Flink version with Kryo v5 support without a bridge version running on Java 11. 
Applications will have to run on a bridge release version that will read their 
state with Kryo v2 data and write it with Kryo v5 data before upgrading to a 
future version of Flink that completely drops support for Kryo v2.

Basically, Flink applications using protobuf dataformat cannot move directly 
from Java 8 to Java 17 without downtime after the kryo v5 release in Flink. 
Applications will need to first move to Java 11 (bridging version) and then 
move to Java 17 to have a safe deployment.

Migration Plan is documented here: 
[https://cwiki.apache.org/confluence/display/FLINK/FLIP-317%3A+Upgrade+Kryo+from+2.24.0+to+5.5.0]

*Blocker for this task:*

Upgrade to Kryo 5.5.0 which supports Java 17 and snapshot migration
https://issues.apache.org/jira/browse/FLINK-3154.


> Move from experimental support to production support for Java 17
> 
>
> Key: FLINK-34491
> URL: https://issues.apache.org/jira/browse/FLINK-34491
> Project: Flink
>  Issue Type: New Feature
>Affects Versions: 1.18.1
>Reporter: Dhruv Patel
>Priority: Major
>
> This task is to move away from experimental support for Java 17 to production 
> support so that teams running Flink in production can migrate to Java 17 
> successfully
> *Background:*
> Flink supports protobuf dataformat to exchange messages between different 
> operators and the serialization and deserialization of those protobufs are 
> performed by library called "Kryo". In order to move away from experimental 
> support of Java 17 released 

[jira] [Updated] (FLINK-34491) Move from experimental support to production support for Java 17

2024-02-28 Thread Dhruv Patel (Jira)


 [ 
https://issues.apache.org/jira/browse/FLINK-34491?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Dhruv Patel updated FLINK-34491:

Description: 
This task is to move away from experimental support for Java 17 to production 
support so that teams running Flink in production can migrate to Java 17 
successfully

*Background:*

Flink supports protobuf dataformat to exchange messages between different 
operators and the serialization and deserialization of those protobufs are 
performed by library called "Kryo". In order to move away from experimental 
support of Java 17 released as part of Flink 1.18.1, the Kryo library in Flink 
1.18.1 needs to be updated from 2.24.0 to 5.5.0 because Kryo 2.24.0 does not 
support Java 17. This improvement plan is tracked as part of this ticket 
https://issues.apache.org/jira/browse/FLINK-3154.

All Flink applications using protobuf currently generate state with Kryo v2. 
Once the above improvement plan is complete all Flink applications will fully 
support reading that state and write newer state with Kryo v5. However, latest 
Kryo v5 doesn't support snapshots made by a previous Kryo v2. This will prevent 
applications which are using snapshot mechanism to deploy their jobs to latest 
Flink version with Kryo v5 support without a bridge version running on Java 11. 
Applications will have to run on a bridge release version that will read their 
state with Kryo v2 data and write it with Kryo v5 data before upgrading to a 
future version of Flink that completely drops support for Kryo v2.

Basically, Flink applications using protobuf dataformat cannot move directly 
from Java 8 to Java 17 without downtime after the kryo v5 release in Flink. 
Applications will need to first move to Java 11 (bridging version) and then 
move to Java 17 to have a safe deployment.

Migration Plan is documented here: 
[https://cwiki.apache.org/confluence/display/FLINK/FLIP-317%3A+Upgrade+Kryo+from+2.24.0+to+5.5.0]

*Blocker for this task:*

Upgrade to Kryo 5.5.0 which supports Java 17 and snapshot migration
https://issues.apache.org/jira/browse/FLINK-3154.

  was:
This task is to move away from experimental support for Java 17 to production 
support so that teams running Flink in production can migrate to Java 17 
successfully

*Background:*

Flink supports protobuf dataformat to exchange messages between different 
operators and the serialization and deserialization of those protobufs are 
performed by library called "Kryo". In order to move away from experimental 
support of Java 17 released as part of Flink 1.18.1, the Kryo library in Flink 
1.18.1 needs to be updated from 2.24.0 to 5.5.0 because Kryo 2.24.0 does not 
support Java 17. This improvement plan is tracked as part of this ticket 
https://issues.apache.org/jira/browse/FLINK-3154.

All Flink applications using protobuf currently generate state with Kryo v2. 
Once the above improvement plan is complete all Flink applications will fully 
support reading that state and write newer state with Kryo v5. However, latest 
Kryo v5 doesn't support snapshots made by a previous Kryo v2. This will prevent 
applications which are using snapshot mechanism to deploy their jobs to latest 
Flink version with Kryo v5 support without a bridge version running on Java 11. 
Applications will have to run on a bridge release version that will read their 
state with Kryo v2 data and write it with Kryo v5 data before upgrading to a 
future version of Flink that completely drops support for Kryo v2.

Basically, Flink applications using protobuf dataformat cannot move directly 
from Java 8 to Java 17 without downtime after the kryo v5 release in Flink. 
Applications will need to first move to Java 11 (bridging version) and then 
move to Java 17 to have a safe deployment.

Migration Plan is documented here: 
https://cwiki.apache.org/confluence/display/FLINK/FLIP-317%3A+Upgrade+Kryo+from+2.24.0+to+5.5.0

*Blocker for this task:*

Upgrade to Kryo 5.5.0 which supports Java 17 
https://issues.apache.org/jira/browse/FLINK-3154.


> Move from experimental support to production support for Java 17
> 
>
> Key: FLINK-34491
> URL: https://issues.apache.org/jira/browse/FLINK-34491
> Project: Flink
>  Issue Type: New Feature
>Affects Versions: 1.18.1
>Reporter: Dhruv Patel
>Priority: Major
>
> This task is to move away from experimental support for Java 17 to production 
> support so that teams running Flink in production can migrate to Java 17 
> successfully
> *Background:*
> Flink supports protobuf dataformat to exchange messages between different 
> operators and the serialization and deserialization of those protobufs are 
> performed by library called "Kryo". In order to move away from experimental 
> support of Java 17 released as part of Flink 1.18.1, the Kryo 

[jira] (FLINK-34491) Move from experimental support to production support for Java 17

2024-02-28 Thread Dhruv Patel (Jira)


[ https://issues.apache.org/jira/browse/FLINK-34491 ]


Dhruv Patel deleted comment on FLINK-34491:
-

was (Author: JIRAUSER289387):
Following issue has been observed in after enabling SSL in flink. Since after 
migration flink uses tls1.3 as default
|Change |Description | |
|SSL / TLS v1.3|the handshake between the flink components now uses TLS v1.3 
with Cipher: 
TLS_AES_256_GCM_SHA384 
which is causing SSL handshake failures. 
 {code:java}
SSL3 alert read:fatal:handshake failure
SSL_connect:error in error
409B7454F87F:error:0A000410:SSL routines:ssl3_read_bytes:sslv3 alert 
handshake failure:ssl/record/rec_layer_s3.c:1586:SSL alert number 40
–
Server Temp Key: ECDH, prime256v1, 256 bits
—
SSL handshake has read 470 bytes and written 730 bytes
Verification: OK
—
New, TLSv1.3, Cipher is TLS_AES_256_GCM_SHA384
This TLS version forbids renegotiation.| {code}| |

> Move from experimental support to production support for Java 17
> 
>
> Key: FLINK-34491
> URL: https://issues.apache.org/jira/browse/FLINK-34491
> Project: Flink
>  Issue Type: New Feature
>Affects Versions: 1.18.1
>Reporter: Dhruv Patel
>Priority: Major
>
> This task is to move away from experimental support for Java 17 to production 
> support so that teams running Flink in production can migrate to Java 17 
> successfully
> *Background:*
> Flink supports protobuf dataformat to exchange messages between different 
> operators and the serialization and deserialization of those protobufs are 
> performed by library called "Kryo". In order to move away from experimental 
> support of Java 17 released as part of Flink 1.18.1, the Kryo library in 
> Flink 1.18.1 needs to be updated from 2.24.0 to 5.5.0 because Kryo 2.24.0 
> does not support Java 17. This improvement plan is tracked as part of this 
> ticket https://issues.apache.org/jira/browse/FLINK-3154.
> All Flink applications using protobuf currently generate state with Kryo v2. 
> Once the above improvement plan is complete all Flink applications will fully 
> support reading that state and write newer state with Kryo v5. However, 
> latest Kryo v5 doesn't support snapshots made by a previous Kryo v2. This 
> will prevent applications which are using snapshot mechanism to deploy their 
> jobs to latest Flink version with Kryo v5 support without a bridge version 
> running on Java 11. Applications will have to run on a bridge release version 
> that will read their state with Kryo v2 data and write it with Kryo v5 data 
> before upgrading to a future version of Flink that completely drops support 
> for Kryo v2.
> Basically, Flink applications using protobuf dataformat cannot move directly 
> from Java 8 to Java 17 without downtime after the kryo v5 release in Flink. 
> Applications will need to first move to Java 11 (bridging version) and then 
> move to Java 17 to have a safe deployment.
> Migration Plan is documented here: 
> https://cwiki.apache.org/confluence/display/FLINK/FLIP-317%3A+Upgrade+Kryo+from+2.24.0+to+5.5.0
> *Blocker for this task:*
> Upgrade to Kryo 5.5.0 which supports Java 17 
> https://issues.apache.org/jira/browse/FLINK-3154.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)


[jira] [Updated] (FLINK-34491) Move from experimental support to production support for Java 17

2024-02-28 Thread Dhruv Patel (Jira)


 [ 
https://issues.apache.org/jira/browse/FLINK-34491?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Dhruv Patel updated FLINK-34491:

Description: 
This task is to move away from experimental support for Java 17 to production 
support so that teams running Flink in production can migrate to Java 17 
successfully

*Background:*

Flink supports protobuf dataformat to exchange messages between different 
operators and the serialization and deserialization of those protobufs are 
performed by library called "Kryo". In order to move away from experimental 
support of Java 17 released as part of Flink 1.18.1, the Kryo library in Flink 
1.18.1 needs to be updated from 2.24.0 to 5.5.0 because Kryo 2.24.0 does not 
support Java 17. This improvement plan is tracked as part of this ticket 
https://issues.apache.org/jira/browse/FLINK-3154.

All Flink applications using protobuf currently generate state with Kryo v2. 
Once the above improvement plan is complete all Flink applications will fully 
support reading that state and write newer state with Kryo v5. However, latest 
Kryo v5 doesn't support snapshots made by a previous Kryo v2. This will prevent 
applications which are using snapshot mechanism to deploy their jobs to latest 
Flink version with Kryo v5 support without a bridge version running on Java 11. 
Applications will have to run on a bridge release version that will read their 
state with Kryo v2 data and write it with Kryo v5 data before upgrading to a 
future version of Flink that completely drops support for Kryo v2.

Basically, Flink applications using protobuf dataformat cannot move directly 
from Java 8 to Java 17 without downtime after the kryo v5 release in Flink. 
Applications will need to first move to Java 11 (bridging version) and then 
move to Java 17 to have a safe deployment.

Migration Plan is documented here: 
https://cwiki.apache.org/confluence/display/FLINK/FLIP-317%3A+Upgrade+Kryo+from+2.24.0+to+5.5.0

*Blocker for this task:*

Upgrade to Kryo 5.5.0 which supports Java 17 
https://issues.apache.org/jira/browse/FLINK-3154.

  was:
This task is to move away from experimental support for Java 17 to production 
support so that teams running Flink in production can migrate to Java 17 
successfully

*Background:*

Flink supports protobuf dataformat to exchange messages between different 
operators and the serialization and deserialization of those protobufs are 
performed by library called "Kryo". In order to move away from experimental 
support of Java 17 released as part of Flink 1.18.1, the Kryo library in Flink 
1.18.1 needs to be updated from 2.24.0 to 5.5.0 because Kryo 2.24.0 does not 
support Java 17. This improvement plan is tracked as part of this ticket 
https://issues.apache.org/jira/browse/FLINK-3154.

All Flink applications using protobuf currently generate state with Kryo v2. 
Once the above improvement plan is complete all Flink applications will fully 
support reading that state and write newer state with Kryo v5. However, latest 
Kryo v5 doesn't support snapshots made by a previous Kryo v2. This will prevent 
applications which are using snapshot mechanism to deploy their jobs to latest 
Flink version with Kryo v5 support without a bridge version running on Java 11. 
Applications will have to run on a bridge release version that will read their 
state with Kryo v2 data and write it with Kryo v5 data before upgrading to a 
future version of Flink that completely drops support for Kryo v2.

Basically, Flink applications using protobuf dataformat cannot move directly 
from Java 8 to Java 17 without downtime. Applications will need to first move 
to Java 11 (bridging version) and then move to Java 17 to have a safe 
deployment.

*Blocker for this task:*

Upgrade to Kryo 5.5.0 which supports Java 17 
https://issues.apache.org/jira/browse/FLINK-3154.


> Move from experimental support to production support for Java 17
> 
>
> Key: FLINK-34491
> URL: https://issues.apache.org/jira/browse/FLINK-34491
> Project: Flink
>  Issue Type: New Feature
>Affects Versions: 1.18.1
>Reporter: Dhruv Patel
>Priority: Major
>
> This task is to move away from experimental support for Java 17 to production 
> support so that teams running Flink in production can migrate to Java 17 
> successfully
> *Background:*
> Flink supports protobuf dataformat to exchange messages between different 
> operators and the serialization and deserialization of those protobufs are 
> performed by library called "Kryo". In order to move away from experimental 
> support of Java 17 released as part of Flink 1.18.1, the Kryo library in 
> Flink 1.18.1 needs to be updated from 2.24.0 to 5.5.0 because Kryo 2.24.0 
> does not support Java 17. This improvement plan is tracked as part of this 
> ticket 

[jira] [Updated] (FLINK-34491) Move from experimental support to production support for Java 17

2024-02-28 Thread Dhruv Patel (Jira)


 [ 
https://issues.apache.org/jira/browse/FLINK-34491?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Dhruv Patel updated FLINK-34491:

Description: 
This task is to move away from experimental support for Java 17 to production 
support so that teams running Flink in production can migrate to Java 17 
successfully

*Background:*

Flink supports protobuf dataformat to exchange messages between different 
operators and the serialization and deserialization of those protobufs are 
performed by library called "Kryo". In order to move away from experimental 
support of Java 17 released as part of Flink 1.18.1, the Kryo library in Flink 
1.18.1 needs to be updated from 2.24.0 to 5.5.0 because Kryo 2.24.0 does not 
support Java 17. This improvement plan is tracked as part of this ticket 
https://issues.apache.org/jira/browse/FLINK-3154.

All Flink applications using protobuf currently generate state with Kryo v2. 
Once the above improvement plan is complete all Flink applications will fully 
support reading that state and write newer state with Kryo v5. However, latest 
Kryo v5 doesn't support snapshots made by a previous Kryo v2. This will prevent 
applications which are using snapshot mechanism to deploy their jobs to latest 
Flink version with Kryo v5 support without a bridge version running on Java 11. 
Applications will have to run on a bridge release version that will read their 
state with Kryo v2 data and write it with Kryo v5 data before upgrading to a 
future version of Flink that completely drops support for Kryo v2.

Basically, Flink applications using protobuf dataformat cannot move directly 
from Java 8 to Java 17 without downtime. Applications will need to first move 
to Java 11 (bridging version) and then move to Java 17 to have a safe 
deployment.

*Blocker for this task:*

Upgrade to Kryo 5.5.0 which supports Java 17 
https://issues.apache.org/jira/browse/FLINK-3154.

  was:
This task is to move away from experimental support for Java 17 to production 
support so that teams running Flink in production can migrate to Java 17 
successfully

*Background:*

Flink supports protobuf dataformat to exchange messages between different 
operators and the serialization and deserialization of those protobufs are 
performed by library called "Kryo". In order to move away from experimental 
support of Java 17 released as part of Flink 1.18.1, the Kryo library in Flink 
1.18.1 needs to be updated from 2.24.0 to 5.5.0 because Kryo 2.24.0 does not 
support Java 17. This improvement plan is tracked as part of this ticket 
https://issues.apache.org/jira/browse/FLINK-3154.

All Flink applications using protobuf currently generate state with Kryo v2. 
Once the above improvement plan is complete all Flink applications will fully 
support reading that state and write newer state with Kryo v5. However, latest 
Kryo v5 doesn't support snapshots made by a previous Kryo v2. This will prevent 
applications which are using snapshot mechanism to deploy their jobs to latest 
Flink version with Kryo v5 support without a bridge version running on Java 11. 
Applications will have to run on a bridge release version that will read their 
state with Kryo v2 data and write it with Kryo v5 data before upgrading to a 
future version of Flink that completely drops support for Kryo v2.

*Blocker for this task:*

Below are some of the blocker task which prevent certain Flink applications 
which use protobuf from upgrading to Java 17
*Blocker 1:* Upgrade to Kryo 5.5.0 which supports Java 17 
https://issues.apache.org/jira/browse/FLINK-3154.

*Blocker 2:* Validate whether a snapshot with Java 8 is compatible with Java 17 
without using a bridge version running on Java 11.
 https://issues.apache.org/jira/browse/FLINK-33707


> Move from experimental support to production support for Java 17
> 
>
> Key: FLINK-34491
> URL: https://issues.apache.org/jira/browse/FLINK-34491
> Project: Flink
>  Issue Type: New Feature
>Affects Versions: 1.18.1
>Reporter: Dhruv Patel
>Priority: Major
>
> This task is to move away from experimental support for Java 17 to production 
> support so that teams running Flink in production can migrate to Java 17 
> successfully
> *Background:*
> Flink supports protobuf dataformat to exchange messages between different 
> operators and the serialization and deserialization of those protobufs are 
> performed by library called "Kryo". In order to move away from experimental 
> support of Java 17 released as part of Flink 1.18.1, the Kryo library in 
> Flink 1.18.1 needs to be updated from 2.24.0 to 5.5.0 because Kryo 2.24.0 
> does not support Java 17. This improvement plan is tracked as part of this 
> ticket https://issues.apache.org/jira/browse/FLINK-3154.
> All Flink applications using protobuf currently generate 

[jira] [Updated] (FLINK-34491) Move from experimental support to production support for Java 17

2024-02-28 Thread Dhruv Patel (Jira)


 [ 
https://issues.apache.org/jira/browse/FLINK-34491?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Dhruv Patel updated FLINK-34491:

Description: 
This task is to move away from experimental support for Java 17 to production 
support so that teams running Flink in production can migrate to Java 17 
successfully

*Background:*

Flink supports protobuf dataformat to exchange messages between different 
operators and the serialization and deserialization of those protobufs are 
performed by library called "Kryo". In order to move away from experimental 
support of Java 17 released as part of Flink 1.18.1, the Kryo library in Flink 
1.18.1 needs to be updated from 2.24.0 to 5.5.0 because Kryo 2.24.0 does not 
support Java 17. This improvement plan is tracked as part of this ticket 
https://issues.apache.org/jira/browse/FLINK-3154.

All Flink applications using protobuf currently generate state with Kryo v2. 
Once the above improvement plan is complete all Flink applications will fully 
support reading that state and write newer state with Kryo v5. However, latest 
Kryo v5 doesn't support snapshots made by a previous Kryo v2. This will prevent 
applications which are using snapshot mechanism to deploy their jobs to latest 
Flink version with Kryo v5 support without a bridge version running on Java 11. 
Applications will have to run on a bridge release version that will read their 
state with Kryo v2 data and write it with Kryo v5 data before upgrading to a 
future version of Flink that completely drops support for Kryo v2.

*Blocker for this task:*

Below are some of the blocker task which prevent certain Flink applications 
which use protobuf from upgrading to Java 17
*Blocker 1:* Upgrade to Kryo 5.5.0 which supports Java 17 
https://issues.apache.org/jira/browse/FLINK-3154.

*Blocker 2:* Validate whether a snapshot with Java 8 is compatible with Java 17 
without using a bridge version running on Java 11.
 https://issues.apache.org/jira/browse/FLINK-33707

  was:
This task is to move away from experimental support for Java 17 to production 
support so that teams running Flink in production can migrate to Java 17 
successfully

 

*Blocker for this task:*

Savepoint migration is not supported with Java 17 and Flink 1.18.1 as mentioned 
in this ticket https://issues.apache.org/jira/browse/FLINK-33707


> Move from experimental support to production support for Java 17
> 
>
> Key: FLINK-34491
> URL: https://issues.apache.org/jira/browse/FLINK-34491
> Project: Flink
>  Issue Type: New Feature
>Affects Versions: 1.18.1
>Reporter: Dhruv Patel
>Priority: Major
>
> This task is to move away from experimental support for Java 17 to production 
> support so that teams running Flink in production can migrate to Java 17 
> successfully
> *Background:*
> Flink supports protobuf dataformat to exchange messages between different 
> operators and the serialization and deserialization of those protobufs are 
> performed by library called "Kryo". In order to move away from experimental 
> support of Java 17 released as part of Flink 1.18.1, the Kryo library in 
> Flink 1.18.1 needs to be updated from 2.24.0 to 5.5.0 because Kryo 2.24.0 
> does not support Java 17. This improvement plan is tracked as part of this 
> ticket https://issues.apache.org/jira/browse/FLINK-3154.
> All Flink applications using protobuf currently generate state with Kryo v2. 
> Once the above improvement plan is complete all Flink applications will fully 
> support reading that state and write newer state with Kryo v5. However, 
> latest Kryo v5 doesn't support snapshots made by a previous Kryo v2. This 
> will prevent applications which are using snapshot mechanism to deploy their 
> jobs to latest Flink version with Kryo v5 support without a bridge version 
> running on Java 11. Applications will have to run on a bridge release version 
> that will read their state with Kryo v2 data and write it with Kryo v5 data 
> before upgrading to a future version of Flink that completely drops support 
> for Kryo v2.
> *Blocker for this task:*
> Below are some of the blocker task which prevent certain Flink applications 
> which use protobuf from upgrading to Java 17
> *Blocker 1:* Upgrade to Kryo 5.5.0 which supports Java 17 
> https://issues.apache.org/jira/browse/FLINK-3154.
> *Blocker 2:* Validate whether a snapshot with Java 8 is compatible with Java 
> 17 without using a bridge version running on Java 11.
>  https://issues.apache.org/jira/browse/FLINK-33707



--
This message was sent by Atlassian Jira
(v8.20.10#820010)


[jira] [Commented] (FLINK-34491) Move from experimental support to production support for Java 17

2024-02-22 Thread Dhruv Patel (Jira)


[ 
https://issues.apache.org/jira/browse/FLINK-34491?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17819860#comment-17819860
 ] 

Dhruv Patel commented on FLINK-34491:
-

Following issue has been observed in after enabling SSL in flink. Since after 
migration flink uses tls1.3 as default
|Change |Description | |
|SSL / TLS v1.3|the handshake between the flink components now uses TLS v1.3 
with Cipher: 
TLS_AES_256_GCM_SHA384 
which is causing SSL handshake failures. 
 {code:java}
SSL3 alert read:fatal:handshake failure
SSL_connect:error in error
409B7454F87F:error:0A000410:SSL routines:ssl3_read_bytes:sslv3 alert 
handshake failure:ssl/record/rec_layer_s3.c:1586:SSL alert number 40
–
Server Temp Key: ECDH, prime256v1, 256 bits
—
SSL handshake has read 470 bytes and written 730 bytes
Verification: OK
—
New, TLSv1.3, Cipher is TLS_AES_256_GCM_SHA384
This TLS version forbids renegotiation.| {code}| |

> Move from experimental support to production support for Java 17
> 
>
> Key: FLINK-34491
> URL: https://issues.apache.org/jira/browse/FLINK-34491
> Project: Flink
>  Issue Type: New Feature
>Affects Versions: 1.18.1
>Reporter: Dhruv Patel
>Priority: Major
>
> This task is to move away from experimental support for Java 17 to production 
> support so that teams running Flink in production can migrate to Java 17 
> successfully
>  
> *Blocker for this task:*
> Savepoint migration is not supported with Java 17 and Flink 1.18.1 as 
> mentioned in this ticket https://issues.apache.org/jira/browse/FLINK-33707



--
This message was sent by Atlassian Jira
(v8.20.10#820010)


[jira] [Updated] (FLINK-34491) Move from experimental support to production support for Java 17

2024-02-22 Thread Dhruv Patel (Jira)


 [ 
https://issues.apache.org/jira/browse/FLINK-34491?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Dhruv Patel updated FLINK-34491:

Description: 
This task is to move away from experimental support for Java 17 to production 
support so that teams running Flink in production can migrate to Java 17 
successfully

 

*Blocker for this task:*

Savepoint migration is not supported with Java 17 and Flink 1.18.1 as mentioned 
in this ticket https://issues.apache.org/jira/browse/FLINK-33707

  was:This task is to move away from experimental support for Java 17 to 
production support so that teams running Flink in production can migrate to 
Java 17 successfully


> Move from experimental support to production support for Java 17
> 
>
> Key: FLINK-34491
> URL: https://issues.apache.org/jira/browse/FLINK-34491
> Project: Flink
>  Issue Type: New Feature
>Affects Versions: 1.18.1
>Reporter: Dhruv Patel
>Priority: Major
>
> This task is to move away from experimental support for Java 17 to production 
> support so that teams running Flink in production can migrate to Java 17 
> successfully
>  
> *Blocker for this task:*
> Savepoint migration is not supported with Java 17 and Flink 1.18.1 as 
> mentioned in this ticket https://issues.apache.org/jira/browse/FLINK-33707



--
This message was sent by Atlassian Jira
(v8.20.10#820010)


[jira] [Updated] (FLINK-34491) Move from experimental support to production support for Java 17

2024-02-21 Thread Dhruv Patel (Jira)


 [ 
https://issues.apache.org/jira/browse/FLINK-34491?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Dhruv Patel updated FLINK-34491:

Description: This task is to move away from experimental support for Java 
17 to production support so that teams running Flink in production can migrate 
to Java 17 successfully  (was: This task is to move away from experimental 
support for Java 17 to production support so that teams running Flink in 
production can migrate to Java 17 successfully

 

https://nightlies.apache.org/flink/flink-docs-master/docs/deployment/java_compatibility/#untested-flink-features-1)

> Move from experimental support to production support for Java 17
> 
>
> Key: FLINK-34491
> URL: https://issues.apache.org/jira/browse/FLINK-34491
> Project: Flink
>  Issue Type: New Feature
>Affects Versions: 1.18.1
>Reporter: Dhruv Patel
>Priority: Major
>
> This task is to move away from experimental support for Java 17 to production 
> support so that teams running Flink in production can migrate to Java 17 
> successfully



--
This message was sent by Atlassian Jira
(v8.20.10#820010)


[jira] [Updated] (FLINK-34491) Move from experimental support to production support for Java 17

2024-02-21 Thread Dhruv Patel (Jira)


 [ 
https://issues.apache.org/jira/browse/FLINK-34491?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Dhruv Patel updated FLINK-34491:

Description: 
This task is to move away from experimental support for Java 17 to production 
support so that teams running Flink in production can migrate to Java 17 
successfully

 

https://nightlies.apache.org/flink/flink-docs-master/docs/deployment/java_compatibility/#untested-flink-features-1

  was:This task is to move away from experimental support for Java 17 to 
production support so that teams running Flink in production can migrate to 
Java 17 successfully


> Move from experimental support to production support for Java 17
> 
>
> Key: FLINK-34491
> URL: https://issues.apache.org/jira/browse/FLINK-34491
> Project: Flink
>  Issue Type: New Feature
>Affects Versions: 1.18.1
>Reporter: Dhruv Patel
>Priority: Major
>
> This task is to move away from experimental support for Java 17 to production 
> support so that teams running Flink in production can migrate to Java 17 
> successfully
>  
> https://nightlies.apache.org/flink/flink-docs-master/docs/deployment/java_compatibility/#untested-flink-features-1



--
This message was sent by Atlassian Jira
(v8.20.10#820010)


[jira] [Updated] (FLINK-34491) Move from experimental support to production support for Java 17

2024-02-21 Thread Dhruv Patel (Jira)


 [ 
https://issues.apache.org/jira/browse/FLINK-34491?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Dhruv Patel updated FLINK-34491:

Description: This task is to move away from experimental support for Java 
17 to production support so that teams running Flink in production can migrate 
to Java 17 successfully  (was: Tracking task to move away from experimental 
support for Java 17 to production support so that teams running Flink in 
production can migrate to Java 17 successfully)

> Move from experimental support to production support for Java 17
> 
>
> Key: FLINK-34491
> URL: https://issues.apache.org/jira/browse/FLINK-34491
> Project: Flink
>  Issue Type: New Feature
>Affects Versions: 1.18.1
>Reporter: Dhruv Patel
>Priority: Major
>
> This task is to move away from experimental support for Java 17 to production 
> support so that teams running Flink in production can migrate to Java 17 
> successfully



--
This message was sent by Atlassian Jira
(v8.20.10#820010)


[jira] [Created] (FLINK-34491) Move from experimental support to production support for Java 17

2024-02-21 Thread Dhruv Patel (Jira)
Dhruv Patel created FLINK-34491:
---

 Summary: Move from experimental support to production support for 
Java 17
 Key: FLINK-34491
 URL: https://issues.apache.org/jira/browse/FLINK-34491
 Project: Flink
  Issue Type: New Feature
Affects Versions: 1.18.1
Reporter: Dhruv Patel


Tracking task to move away from experimental support for Java 17 to production 
support so that teams running Flink in production can migrate to Java 17 
successfully



--
This message was sent by Atlassian Jira
(v8.20.10#820010)


[jira] [Updated] (FLINK-32015) Flink does not fail SSL Handshake when expired cert is used

2023-05-05 Thread Dhruv Patel (Jira)


 [ 
https://issues.apache.org/jira/browse/FLINK-32015?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Dhruv Patel updated FLINK-32015:

Description: 
We recently observed that Flink is able to perform SSL communication without 
any issues even if certs in the trust store are expired

As per this issue in Netty library, it says that Default TrustManagerFactory 
does not come with that check in place and we need to implement our own 
TrustManagerFactory for the desired behavior
[https://github.com/netty/netty/issues/8461]

It would be good to perform the cert validity check to prevent the risk of 
using expired certs

  was:
We recently observed that Flink is able to perform SSL communication without 
any issues even if certs in the trust store are expired

As per this issue in Netty library, it says that Default TrustManagerFactory 
does not come with that check in place and we need to implement our own 
TrustManagerFactory for the desired behavior
https://github.com/netty/netty/issues/8461


> Flink does not fail SSL Handshake when expired cert is used
> ---
>
> Key: FLINK-32015
> URL: https://issues.apache.org/jira/browse/FLINK-32015
> Project: Flink
>  Issue Type: Bug
>  Components: Runtime / Network
>Reporter: Dhruv Patel
>Priority: Major
>
> We recently observed that Flink is able to perform SSL communication without 
> any issues even if certs in the trust store are expired
> As per this issue in Netty library, it says that Default TrustManagerFactory 
> does not come with that check in place and we need to implement our own 
> TrustManagerFactory for the desired behavior
> [https://github.com/netty/netty/issues/8461]
> It would be good to perform the cert validity check to prevent the risk of 
> using expired certs



--
This message was sent by Atlassian Jira
(v8.20.10#820010)


[jira] [Created] (FLINK-32015) Flink does not fail SSL Handshake when expired cert is used

2023-05-05 Thread Dhruv Patel (Jira)
Dhruv Patel created FLINK-32015:
---

 Summary: Flink does not fail SSL Handshake when expired cert is 
used
 Key: FLINK-32015
 URL: https://issues.apache.org/jira/browse/FLINK-32015
 Project: Flink
  Issue Type: Bug
  Components: Runtime / Network
Reporter: Dhruv Patel


We recently observed that Flink is able to perform SSL communication without 
any issues even if certs in the trust store are expired

As per this issue in Netty library, it says that Default TrustManagerFactory 
does not come with that check in place and we need to implement our own 
TrustManagerFactory for the desired behavior
https://github.com/netty/netty/issues/8461



--
This message was sent by Atlassian Jira
(v8.20.10#820010)


[jira] [Created] (FLINK-29400) Default Value of env.log.max in documentation is incorrect

2022-09-22 Thread Dhruv Patel (Jira)
Dhruv Patel created FLINK-29400:
---

 Summary: Default Value of env.log.max in documentation is incorrect
 Key: FLINK-29400
 URL: https://issues.apache.org/jira/browse/FLINK-29400
 Project: Flink
  Issue Type: Improvement
  Components: Documentation
Reporter: Dhruv Patel


The default value of env.log.max is 10 as per the code in master 
([https://github.com/apache/flink/blob/master/flink-dist/src/main/flink-bin/bin/config.sh#L137).]

However the Flink Documentation says the default value is 5 
(https://nightlies.apache.org/flink/flink-docs-master/docs/deployment/config/#env-log-max)
 which is incorrect 



--
This message was sent by Atlassian Jira
(v8.20.10#820010)


[jira] [Created] (FLINK-27590) Provide endpoint with information about leader Jobmanager in HA setups via Zookeeper

2022-05-12 Thread Dhruv Patel (Jira)
Dhruv Patel created FLINK-27590:
---

 Summary: Provide endpoint with information about leader Jobmanager 
in HA setups via Zookeeper
 Key: FLINK-27590
 URL: https://issues.apache.org/jira/browse/FLINK-27590
 Project: Flink
  Issue Type: Improvement
  Components: Runtime / REST
Affects Versions: 1.15.0, 1.14.0
Reporter: Dhruv Patel
 Attachments: error.log

In High Availability setups via Zookeeper, it would be nice to provide 
information about the leader jobmanager. Although, the job submission and all 
the operations can be performed via any jobmanager. It would be nice to expose 
data about leader jobmanager due to following reasons
 * Since there are multiple job managers in HA setup, each of them will expose 
a web ui and a rest endpoint. It is difficult to know which of the endpoint is 
backed by active jobmanager. Moreover, we cannot access the taskmanager logs 
from the flink ui exposed by the standby jobmanager. [^error.log]
 * If the flink UI/REST endpoint is fronted by a Load Balancer, it would be 
nice to detect active jobmanager so that all the UI requests can be routed to 
the active jobmanager to avoid having issues in accessing logs



--
This message was sent by Atlassian Jira
(v8.20.7#820007)