Uh oh!
There was an error while loading. Please reload this page.
HBASE-28637 asyncwal should attempt to recover lease if close fails - #5962
Conversation
Change-Id: I08dc88c8d6ba4e23754ef2acd1e14b32b225ae17
Apache-HBase
commented
Jun 4, 2024
💔 -1 overall
This message was automatically generated. |
Apache-HBase
commented
Jun 4, 2024
🎊 +1 overall
This message was automatically generated. |
Apache-HBase
commented
Jun 5, 2024
🎊 +1 overall
This message was automatically generated. |
Apache-HBase
commented
Jun 5, 2024
🎊 +1 overall
This message was automatically generated. |
Apache9
commented
Jun 5, 2024
FSHLog does not need this logic? |
| try { | ||
| RecoverLeaseFSUtils.recoverFileLease(fs, p, conf, null); | ||
| } catch (IOException ex) { | ||
| LOG.warn("Unable to recover lease after several attempts. Give up.", ex); |
Change-Id: Id2906597d4709d66db370886598f339cd77ee754
jojochuang
commented
Jun 5, 2024
@Apache9 updated thanks for the review. |
Apache-HBase
commented
Jun 5, 2024
💔 -1 overall
This message was automatically generated. |
Apache-HBase
commented
Jun 5, 2024
🎊 +1 overall
This message was automatically generated. |
Apache-HBase
commented
Jun 5, 2024
💔 -1 overall
This message was automatically generated. |
Apache-HBase
commented
Jun 5, 2024
🎊 +1 overall
This message was automatically generated. |
Apache-HBase
commented
Jun 7, 2024
🎊 +1 overall
This message was automatically generated. |
Apache-HBase
commented
Jun 7, 2024
💔 -1 overall
This message was automatically generated. |
Apache-HBase
commented
Jun 7, 2024
🎊 +1 overall
This message was automatically generated. |
Apache-HBase
commented
Jun 7, 2024
🎊 +1 overall
This message was automatically generated. |
| RecoverLeaseFSUtils.recoverFileLease(fs, p, conf, null); | ||
| } catch (IOException ex) { | ||
| LOG.error("Unable to recover lease after several attempts. Give up.", ex); | ||
| throw new RuntimeException(ex); |
There was a problem hiding this comment.
Better not do this...
We are running in a background thread pool and this will kill the thread in the thread pool...
If the thread pool can still create new threads then we are safe, but it is useless to throw the exception out here. And if the thread pool is broken, then we are in trouble...
There was a problem hiding this comment.
Good point. How about just log the exception and move on?
Change-Id: I68d3b6e15c141433321c58b66b8de9efb10b171d
Apache-HBase
commented
Jun 10, 2024
🎊 +1 overall
This message was automatically generated. |
Apache-HBase
commented
Jun 10, 2024
💔 -1 overall
This message was automatically generated. |
Apache-HBase
commented
Jun 10, 2024
🎊 +1 overall
This message was automatically generated. |
Apache-HBase
commented
Jun 10, 2024
🎊 +1 overall
This message was automatically generated. |
Apache-HBase
commented
Jun 11, 2024
🎊 +1 overall
This message was automatically generated. |
Apache-HBase
commented
Jun 11, 2024
🎊 +1 overall
This message was automatically generated. |
Apache-HBase
commented
Jun 11, 2024
🎊 +1 overall
This message was automatically generated. |
Apache-HBase
commented
Jun 11, 2024
🎊 +1 overall
This message was automatically generated. |
jojochuang
commented
Jun 12, 2024
Thanks @Apache9 ! |
…pache#5962) Signed-off-by: Duo Zhang <zhangduo@apache.org> (cherry picked from commit 0e8cfdb) (cherry picked from commit b881296) Conflicts: hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/wal/AbstractFSWAL.java Change-Id: Ia704c1b36b99be345d11a70ac2452dea094e9575
…lose fails (apache#5962) (apache#6033) Signed-off-by: Duo Zhang <zhangduo@apache.org> (cherry picked from commit 0e8cfdb) (cherry picked from commit b881296) Change-Id: I6ad1d3122a0452d583a7269be9dbc3851f4f3a37
…lose fails (apache#5962) (apache#6033) Signed-off-by: Duo Zhang <zhangduo@apache.org> (cherry picked from commit 0e8cfdb) (cherry picked from commit b881296) Change-Id: I6ad1d3122a0452d583a7269be9dbc3851f4f3a37 (cherry picked from commit b68fecf)
…lose fails (apache#5962) (apache#6033) Signed-off-by: Duo Zhang <zhangduo@apache.org> (cherry picked from commit 0e8cfdb) (cherry picked from commit b881296) Change-Id: I6ad1d3122a0452d583a7269be9dbc3851f4f3a37 (cherry picked from commit b68fecf)
Do lease recovery as a last resort if async wal fails to close WAL file.
This issue will make WAL on Ozone more resilient because Ozone does not support renaming open files (which happens due to WAL log archiving).