-
Notifications
You must be signed in to change notification settings - Fork 3.3k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
HBASE-28900: Avoid resetting the bucket cache during recovery from persistence. #6360
Conversation
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
"Couldn't find match for index " + idx + " in free list"); | ||
if (!matchFound) { | ||
LOG.warn("We found more entries for bucket than the bucket capacity. " | ||
+ "Skipping entry at cache offset {}", offset); |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
nit: let's add more info about the bucket. Something like:
LOG.warn("We found more entries for bucket starting at offset {} for blocks of {} size. "
+ "Skipping entry at cache offset {}", baseOffset, itemAllocationSize, offset);
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
ack!
…rsistence. When an inconsistency is detected during the recovery of bucket cache from persistence, we tend to throw away the complete cache and try to rebuild the cache. Thsi inconsistency can occur when the region server is abruptly terminated. This can be avoided by skipping the inconsistent backing map entry and continuing with the cache recovery. The inconsistent backing map entry will be discarded during the subsequent bucket cache validation. Change-Id: I41237bda2189cbd73e22e4eadf4d53d16a3733f6
🎊 +1 overall
This message was automatically generated. |
🎊 +1 overall
This message was automatically generated. |
…rsistence. (#6360) Signed-off-by: Wellington Chevreuil <wchevreuil@apache.org>
…rsistence. (apache#6360) Signed-off-by: Wellington Chevreuil <wchevreuil@apache.org>
…rsistence. (apache#6360) Signed-off-by: Wellington Chevreuil <wchevreuil@apache.org> Change-Id: If7e7a76723cf7d2d6e8c9e2ce0aeccfdba81c0bb
When an inconsistency is detected during the recovery of bucket cache from persistence, we tend to throw away the complete cache and try to rebuild the cache. Thsi inconsistency can occur when the region server is abruptly terminated.
This can be avoided by skipping the inconsistent backing map entry and continuing with the cache recovery. The inconsistent backing map entry will be discarded during the subsequent bucket cache validation.
Change-Id: I41237bda2189cbd73e22e4eadf4d53d16a3733f6