Update defaults of max-concurrent-file-group-rewrites to 5

[karuppayya]

No description provided.

[karuppayya]

@aokolnychyi @RussellSpitzer @flyrain @szehon-ho

[aokolnychyi]

I support the idea of changing this default as we ask our users to change it in 9/10 cases.
@RussellSpitzer, any thoughts on this as you wrote this part?

[RussellSpitzer]

I like 5 but what I'd really like us to do is check the size of our "file groups" then use that to determine what to set max concurrent groups too. For example if we see all file groups are less than say target file size, we set Max Conncurrent = Num Cores

[aokolnychyi]

We would probably need to look at the scheduler implementation (e.g. FIFO) and total resources available in the cluster vs the amount of data we need to compact and partitions. I think it makes sense exploring but if it becomes too tricky, I'd just default it to something more reasonable.

[ajantha-bhat]

I support the idea of changing this default as we ask our users to change it in 9/10 cases.

True. Weekly once or twice, I am suggesting users configure it in Iceberg slack. Most of them doesn't read javadoc :(
I would love to see this default value getting changed.

+1 for this change.

[RussellSpitzer]

I'm sold for this right now, then something more complicated later

[ajantha-bhat]

It seems now testPartialProgressEnabled is flaky. failed only for spark3.1 java 8.
Because of 5 concurrent commits, One of the file groups has failed to commit? I think it is better to set the concurrency to 1 for this particular test case.

org.apache.iceberg.spark.actions.TestRewriteDataFilesAction > testPartialProgressEnabled FAILED
    java.lang.AssertionError: Should have 10 fileGroups expected:<9> but was:<10>
        at org.junit.Assert.fail(Assert.java:89)
        at org.junit.Assert.failNotEquals(Assert.java:835)
        at org.junit.Assert.assertEquals(Assert.java:647)
        at 

[karuppayya]

The test was failing as one(or more) commit was failing

[Rewrite-Service-1] ERROR org.apache.iceberg.actions.RewriteDataFilesCommitManager - Cannot commit groups [RewriteFileGroup{info=BaseRewriteDataFilesFileGroupInfo{globalIndex=2, partitionIndex=2, 

This was happening since the number of concurrent writes was more than the retry attempts.
I will fix the test, to increase the retry attempts.
I will also file an issue to see if this can be computed and passed on to action.
@RussellSpitzer @aokolnychyi @ajantha-bhat

[RussellSpitzer]

Thanks for this update @karuppayya !

[nastra]

typo: casse -> cases

[nastra]

I think this should land under 1.2.0 and not 1.1.0 because the change is being introduced after 1.2.0 was released (most likely because the branch was created before 1.2.0 was released). @karuppayya could you please follow-up with that and move this under 1.2.0?

[RussellSpitzer]

Ah yeah I forgot we just missed the release

[nastra]

I've opened #7223 to address that. @RussellSpitzer could you review please?