Re: [PR] HDFS-17497. The number of bytes of the last committed block should be calculated into the file length [hadoop]

haiyang1987 commented on code in PR #6765: URL: https://github.com/apache/hadoop/pull/6765#discussion_r1599287673 ## hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/BlockInfo.java: ## @@ -387,6 +387,19 @@ public boolean

Re: [PR] HDFS-17497. The number of bytes of the last committed block should be calculated into the file length [hadoop]

hfutatzhanghb commented on code in PR #6765: URL: https://github.com/apache/hadoop/pull/6765#discussion_r1597392676 ## hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/namenode/FSDirectory.java: ## @@ -1105,15 +1106,12 @@ static void

Re: [PR] HDFS-17497. The number of bytes of the last committed block should be calculated into the file length [hadoop]

ZanderXu commented on PR #6765: URL: https://github.com/apache/hadoop/pull/6765#issuecomment-2081766984 > IIRC, client also check the file length through request DataNode which manage the uncomplete block? Like other committed blocks, the client does not need to get the visible

Re: [PR] HDFS-17497. The number of bytes of the last committed block should be calculated into the file length [hadoop]

hadoop-yetus commented on PR #6765: URL: https://github.com/apache/hadoop/pull/6765#issuecomment-2081636005 :confetti_ball: **+1 overall** | Vote | Subsystem | Runtime | Logfile | Comment | |::|--:|:|::|:---:| _ Prechecks

Re: [PR] HDFS-17497. The number of bytes of the last committed block should be calculated into the file length [hadoop]

Hexiaoqiao commented on PR #6765: URL: https://github.com/apache/hadoop/pull/6765#issuecomment-2081489723 Great catch, not review carefully, but I remember this have been discussed for long time. IIRC, client also check the file length through request DataNode which manage the uncomplete

Re: [PR] HDFS-17497. The number of bytes of the last committed block should be calculated into the file length [hadoop]

hadoop-yetus commented on PR #6765: URL: https://github.com/apache/hadoop/pull/6765#issuecomment-2074475186 :confetti_ball: **+1 overall** | Vote | Subsystem | Runtime | Logfile | Comment | |::|--:|:|::|:---:| | +0 :ok: |

Re: [PR] HDFS-17497. The number of bytes of the last committed block should be calculated into the file length [hadoop]

ZanderXu commented on PR #6765: URL: https://github.com/apache/hadoop/pull/6765#issuecomment-2073863664 @xkrogen master, I very much hope you can review this PR when you are available, since you are familiar with ~~HDFS-10843~~ -- This is an automated message from the Apache Git Service.

Re: [PR] HDFS-17497. The number of bytes of the last committed block should be calculated into the file length [hadoop]

hadoop-yetus commented on PR #6765: URL: https://github.com/apache/hadoop/pull/6765#issuecomment-2073379567 :broken_heart: **-1 overall** | Vote | Subsystem | Runtime | Logfile | Comment | |::|--:|:|::|:---:| | +0 :ok: |

[PR] HDFS-17497. The number of bytes of the last committed block should be calculated into the file length [hadoop]

ZanderXu opened a new pull request, #6765: URL: https://github.com/apache/hadoop/pull/6765 One in-writing HDFS file may contains multiple committed blocks, as follows (assume one file contains three blocks):   | Block 1 | Block 2 | Block 3 -- | -- | -- | -- Case 1 |