[#6307] feat(flink): remove log4j from Gravitino Flink connector

[FANNG1]

What changes were proposed in this pull request?

remove log4j from Gravitino Flink connector

Why are the changes needed?

Fix: #6307

Does this PR introduce any user-facing change?

no

How was this patch tested?

setup an local flink cluster, run flink SQL

[FANNG1]

@coolderli @jerryshao PTAL

[yuqi1129]

Have you identified the root cause of this issue? Does it work with 0.7.0-inbucating?

[FANNG1]

Have you identified the root cause of this issue? Does it work with 0.7.0-inbucating?

yes, it works with 0.7.0-inbucating

[yuqi1129]

Have you identified the root cause of this issue? Does it work with 0.7.0-inbucating?

yes, it works with 0.7.0-inbucating

So I suggest we may need to find out which commit introduced the problem and make the necessary changes accordingly.

[FANNG1]

Have you identified the root cause of this issue? Does it work with 0.7.0-inbucating?

yes, it works with 0.7.0-inbucating

So I suggest we may need to find out which commit introduced the problem and make the necessary changes accordingly.

I don' think so, it's reasonable to remove log4j from the Flink runtime jar to avoid the conflict.

[yuqi1129]

Have you identified the root cause of this issue? Does it work with 0.7.0-inbucating?

yes, it works with 0.7.0-inbucating

So I suggest we may need to find out which commit introduced the problem and make the necessary changes accordingly.

I don' think so, it's reasonable to remove log4j from the Flink runtime jar to avoid the conflict.

I don't deny this remove log4j from the Flink runtime jar is resonable, what I emphasize is that we should know which commit introduces this problem.

[FANNG1]

Have you identified the root cause of this issue? Does it work with 0.7.0-inbucating?

yes, it works with 0.7.0-inbucating

So I suggest we may need to find out which commit introduced the problem and make the necessary changes accordingly.

I don' think so, it's reasonable to remove log4j from the Flink runtime jar to avoid the conflict.

I don't deny this remove log4j from the Flink runtime jar is resonable, what I emphasize is that we should know which commit introduces this problem.

we didn't upgrade log4j version for one year, seems not the commit in Gravitino introduce this problems. I guess the new package changes the sequence of the classes loaded by Flink classloader introduce this problem.