Uh oh!
There was an error while loading. Please reload this page.
HBASE-27238 Backport backup restore to 2.x - #4770
Merged
bbeaudreault merged 4 commits intoJan 24, 2023
Merged
Conversation
Apache-HBase
commented
Sep 8, 2022
💔 -1 overall
This message was automatically generated. |
Apache-HBase
commented
Sep 8, 2022
💔 -1 overall
This message was automatically generated. |
Apache-HBase
commented
Sep 8, 2022
💔 -1 overall
This message was automatically generated. |
rda3monforce-pushed
the
HBASE-27238-backport-backup-restore-to-2.x
branch
from
September 8, 2022 06:07
7dba1c5 to
7350663CompareApache-HBase
commented
Sep 8, 2022
🎊 +1 overall
This message was automatically generated. |
Apache-HBase
commented
Sep 8, 2022
💔 -1 overall
This message was automatically generated. |
Apache-HBase
commented
Sep 12, 2022
💔 -1 overall
This message was automatically generated. |
Apache-HBase
commented
Sep 12, 2022
🎊 +1 overall
This message was automatically generated. |
Apache-HBase
commented
Sep 12, 2022
🎊 +1 overall
This message was automatically generated. |
bbeaudreault
commented
Sep 12, 2022
Contributor
@rda3mon I know this is still just a draft, but can you address the error-prone findings in the pre-commit? Shows up under javac: https://ci-hbase.apache.org/job/HBase-PreCommit-GitHub-PR/job/PR-4770/3/artifact/yetus-general-check/output/diff-compile-javac-root.txt Can you do it in a separate commit so that I can review it separately from the initial backport? I re-ran the pre-commit hooks and all the tests are passing, the above is the only automated failure test remaining. The spotbugs error is a known issue which I think will be fixed in https://issues.apache.org/jira/browse/HBASE-27363 |
rda3mon
commented
Sep 13, 2022
ContributorAuthor
@bbeaudreault These errors if I fix, would differ from master branch. right? |
rda3mon
commented
Sep 13, 2022
ContributorAuthor
@bbeaudreault Is there a way to run these checks locally? Pushing and checking might be tedious job to do for fixing non obvious ones. |
bbeaudreault
commented
Sep 13, 2022
Contributor
@rda3mon I find that the violations are usually pretty straightforward. If there are any that you have a question about, feel free to post here. I do think it's possible to run locally, maybe with But actually, you could try including 2c3abae in your back port here. Seems like Andrew already handled these in master. This has me wondering if there are other patches we need to include though. |
ContributorAuthor
Highly unlikely. I did go into each of the commits from backup history to verify if anything is missed out. Looks like this pr was merged after that. |
Apache-HBase
commented
Sep 13, 2022
💔 -1 overall
This message was automatically generated. |
Apache-HBase
commented
Sep 13, 2022
💔 -1 overall
This message was automatically generated. |
Apache-HBase
commented
Sep 13, 2022
💔 -1 overall
This message was automatically generated. |
Apache-HBase
commented
Sep 13, 2022
💔 -1 overall
This message was automatically generated. |
Apache-HBase
commented
Sep 13, 2022
🎊 +1 overall
This message was automatically generated. |
Apache-HBase
commented
Sep 13, 2022
💔 -1 overall
This message was automatically generated. |
bbeaudreault
commented
Sep 13, 2022
Contributor
@rda3mon looks like that cherry-pick resolved most of the error prone issues. Looks like there are 3 left: https://ci-hbase.apache.org/job/HBase-PreCommit-GitHub-PR/job/PR-4770/5/artifact/yetus-general-check/output/diff-compile-javac-root.txt They look pretty simple/javadoc related. would you mind pushing 1 more commit to fix those. otherwise the build looks clean; the test failure looks unrelated. once you push that fix, i can re-run the tests if needed. Gonna start reviewing the code this week hopefully. Can you link me to specific areas of this PR that had merge conflicts or net-new code, if possible. I know in slack you mentioned WALPlayer; I'll take a look at that. Any others would be helpful |
rda3monforce-pushed
the
HBASE-27238-backport-backup-restore-to-2.x
branch
2 times, most recently
from
September 13, 2022 14:34
9d89e43 to
5f92b64CompareApache-HBase
commented
Sep 13, 2022
💔 -1 overall
This message was automatically generated. |
Apache-HBase
commented
Sep 13, 2022
🎊 +1 overall
This message was automatically generated. |
Apache-HBase
commented
Sep 13, 2022
🎊 +1 overall
This message was automatically generated. |
Apache-HBase
commented
Sep 13, 2022
💔 -1 overall
This message was automatically generated. |
Apache-HBase
commented
Sep 14, 2022
🎊 +1 overall
This message was automatically generated. |
Apache-HBase
commented
Sep 14, 2022
🎊 +1 overall
This message was automatically generated. |
rda3mon
marked this pull request as ready for review
September 14, 2022 06:13
Apache-HBase
commented
Jan 19, 2023
🎊 +1 overall
This message was automatically generated. |
rda3monforce-pushed
the
HBASE-27238-backport-backup-restore-to-2.x
branch
from
January 20, 2023 02:40
4ccda54 to
db06805CompareApache-HBase
commented
Jan 20, 2023
🎊 +1 overall
This message was automatically generated. |
rda3monforce-pushed
the
HBASE-27238-backport-backup-restore-to-2.x
branch
from
January 20, 2023 04:24
db06805 to
7155d53CompareApache-HBase
commented
Jan 20, 2023
💔 -1 overall
This message was automatically generated. |
Apache-HBase
commented
Jan 20, 2023
🎊 +1 overall
This message was automatically generated. |
Apache-HBase
commented
Jan 20, 2023
🎊 +1 overall
This message was automatically generated. |
bbeaudreault
commented
Jan 20, 2023
Contributor
Nice! I think this looks good. Just waiting on clean pre-commit checks. |
Contributor
There are some javac (errorprone) warnings. They are all in the net-new code, and I checked them and master has the same issues. Seems like we skipped cleaning up errorprone in backup/restore prior to now. Let's ignore those failures, I think I will create a separate jira to do errorprone cleanup of edit: tracked in https://issues.apache.org/jira/browse/HBASE-27582 |
Apache-HBase
commented
Jan 20, 2023
🎊 +1 overall
This message was automatically generated. |
Apache-HBase
commented
Jan 20, 2023
🎊 +1 overall
This message was automatically generated. |
rda3mon
commented
Jan 21, 2023
ContributorAuthor
@bbeaudreault How to re run the CI pipeline? |
bbeaudreault
left a comment
Contributor
There was a problem hiding this comment.
sorry @rda3mon -- a couple more small changes. We really are so close here, i just missed the nuance in WALPlayer here. But this is the last class with modifications, then we can merge.
Comment on lines
+119
to
+123
| byte[] outKey = multiTableSupport | ||
| ? Bytes.add(table.getName(), Bytes.toBytes(tableSeparator), | ||
| CellUtil.cloneRow(KeyValueUtil.ensureKeyValue(cell))) | ||
| : CellUtil.cloneRow(KeyValueUtil.ensureKeyValue(cell)); | ||
| context.write(new ImmutableBytesWritable(outKey), KeyValueUtil.ensureKeyValue(cell)); |
Contributor
There was a problem hiding this comment.
I fear this block can cause a performance regression. Taking a look at KeyValueUtil.ensureKeyValue, depending on the type of cell, ensureKeyValue can result in the creation of a new KeyValue object. Unnecessary object creation in hadoop jobs can be expensive over a long job.
Can you please update this to only call KeyValueUtil.ensureKeyValue once for each cell?
ContributorAuthor
There was a problem hiding this comment.
Done. A miss from my side.
Comment on lines
+83
to
+89
| // This relies on Hadoop Configuration to handle warning about deprecated configs and | ||
| // to set the correct non-deprecated configs when an old one shows up. | ||
| static { | ||
| Configuration.addDeprecation("hlog.bulk.output", BULK_OUTPUT_CONF_KEY); | ||
| Configuration.addDeprecation("hlog.input.tables", TABLES_KEY); | ||
| Configuration.addDeprecation("hlog.input.tablesmap", TABLE_MAP_KEY); | ||
| } |
Contributor
There was a problem hiding this comment.
Was there a reason to add these? They aren't an issue per-se, but I noticed these deprecations are already in HBaseConfiguration class.
ContributorAuthor
There was a problem hiding this comment.
Looks like unintended changes. I will remove these. Must have come as part of some commit cherry-picking. And they are already present as part of HbaseConfiguration as mentioned.
Comment on lines
+367
to
+374
| try (Connection conn = ConnectionFactory.createConnection(conf);) { | ||
| List<TableInfo> tableInfoList = new ArrayList<>(); | ||
| for (TableName tableName : tableNames) { | ||
| Table table = conn.getTable(tableName); | ||
| RegionLocator regionLocator = conn.getRegionLocator(tableName); | ||
| tableInfoList.add(new TableInfo(table.getDescriptor(), regionLocator)); | ||
| } | ||
| MultiTableHFileOutputFormat.configureIncrementalLoad(job, tableInfoList); |
Contributor
There was a problem hiding this comment.
Ok so I think we might want to make one more compatibility fix here:
Previously, if a user called this method with more than 1 table it would throw an exception.
Now, those tables will be collected and passed into MultiTableHFileOutputFormat.
That is fine I think. However, if you call this method with just 1 table (as most users would do today since > 1 throws an error), you also will go to MultiTableHFileOutputFormat.
I think on line 374 here we should just be defensive and do this:
if (tableInfoList.size() > 1) {
MultiTableHFileOutputFormat.configureIncrementalLoad(job, tableInfoList);
} else {
TableInfo tableInfo = tableInfoList.get(0);
HFileOutputFormat2.configureIncrementalLoad(job, tableInfo.getTableDescriptor(), tableInfo.getRegionLocator());
}
This is important because if you look at HFileOutputFormat2.configureIncrementalLoad there is this:
boolean writeMultipleTables = false;
if (MultiTableHFileOutputFormat.class.equals(cls)) {
writeMultipleTables = true;
conf.setBoolean(MULTI_TABLE_HFILEOUTPUTFORMAT_CONF_KEY, true);
}
So the PR as we have it here would cause all users of WALPlayer to automatically pick up the writeMultipleTables behavior, even if they just use 1 table. The writeMultipleTables behavior involves a bunch of differences from partitioner, splits, output directory, etc.
My suggestion above would ensure that you only go into the writeMultipleTable mode if you're actually writing to multiple tables as indicated by tableInfoList.size() > 1.
bbeaudreault
commented
Jan 21, 2023
Contributor
Only committers can do it. Actually I was going to merge anyway, cuz we got 3 +1's which is all we really need. The pipeline failed on the last reporting step, unrelated. But now just the few more comments above |
Signed-off-by: Duo Zhang <zhangduo@apache.org> Signed-off-by: Viraj Jasani <vjasani@apache.org>
rda3monforce-pushed
the
HBASE-27238-backport-backup-restore-to-2.x
branch
from
January 22, 2023 10:06
7155d53 to
fc898a7CompareApache-HBase
commented
Jan 22, 2023
🎊 +1 overall
This message was automatically generated. |
Apache-HBase
commented
Jan 22, 2023
💔 -1 overall
This message was automatically generated. |
rda3monforce-pushed
the
HBASE-27238-backport-backup-restore-to-2.x
branch
from
January 23, 2023 17:15
fc898a7 to
2bf4f6aCompareApache-HBase
commented
Jan 23, 2023
🎊 +1 overall
This message was automatically generated. |
Apache-HBase
commented
Jan 23, 2023
🎊 +1 overall
This message was automatically generated. |
Apache-HBase
commented
Jan 24, 2023
🎊 +1 overall
This message was automatically generated. |
bbeaudreault
approved these changes
Jan 24, 2023
bbeaudreault
left a comment
Contributor
There was a problem hiding this comment.
All changes look good. We've retained backwards compatibility with 2.x releases, no expected changes for existing users from normal behavior. They can enable hbase.hfileoutputformat.tablename.namespace.inclusive to trigger change in directory structure to disambiguate tables in different namespaces, which will be default in 3.0. This is enabled by default for backups, as is wal.multi.tables.support which enables MultiTableHFileOutputFormat for WALPlayer bulkload mode. Otherwise, users can manually enable wal.multi.tables.support in WALPlayer if they want that behavior themselves.
All tests are passing, all net-new files confirmed equivalent with master. Note one small difference from master in IncrementalTableBackupClient.java, which adds handling of our compatiblity configs.
Final build step failed due to no space left on the jenkins host, but we got 3 +1's indicating success for this PR.
Thanks @rda3mon!
rda3mon
commented
Jan 24, 2023
ContributorAuthor
@bbeaudreault Thanks for taking time out to review this lengthy PR. |
bbeaudreault pushed a commit
to HubSpot/hbase
that referenced
this pull request
Jan 24, 2023
…4770) Signed-off-by: Bryan Beaudreault <bbeaudreault@apache.org>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for freeto join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
No description provided.