[jira] [Comment Edited] (HBASE-13877) Interrupt to flush from TableFlushProcedure causes dataloss in ITBLL
[ https://issues.apache.org/jira/browse/HBASE-13877?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14579764#comment-14579764 ] Enis Soztutar edited comment on HBASE-13877 at 6/10/15 12:06 AM: - bq. There is no other caller? Just call shutdown rather than shutdownNow? Yeah, looked like the intention for writing that method was to do interruption, but it is called from cancelTasks() which calls cancel(false). Let me update the patch. was (Author: enis): bq. There is no other caller? Just call shutdown rather than shutdownNow? Yeah, looked like the intention for writing that method to do interruption, but it is called from cancelTasks() which calls cancel(false). Let me update the patch. Interrupt to flush from TableFlushProcedure causes dataloss in ITBLL Key: HBASE-13877 URL: https://issues.apache.org/jira/browse/HBASE-13877 Project: HBase Issue Type: Bug Reporter: Enis Soztutar Assignee: Enis Soztutar Priority: Blocker Fix For: 2.0.0, 1.2.0, 1.1.1 Attachments: hbase-13877_v1.patch, hbase-13877_v2-branch-1.1.patch ITBLL with 1.25B rows failed for me (and Stack as reported in https://issues.apache.org/jira/browse/HBASE-13811?focusedCommentId=14577834page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-14577834) HBASE-13811 and HBASE-13853 fixed an issue with WAL edit filtering. The root cause this time seems to be different. It is due to procedure based flush interrupting the flush request in case the procedure is cancelled from an exception elsewhere. This leaves the memstore snapshot intact without aborting the server. The next flush, then flushes the previous memstore with the current seqId (as opposed to seqId from the memstore snapshot). This creates an hfile with larger seqId than what its contents are. Previous behavior in 0.98 and 1.0 (I believe) is that after flush prepare and interruption / exception will cause RS abort. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Comment Edited] (HBASE-13877) Interrupt to flush from TableFlushProcedure causes dataloss in ITBLL
[
https://issues.apache.org/jira/browse/HBASE-13877?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14579870#comment-14579870
]
Duo Zhang edited comment on HBASE-13877 at 6/10/15 2:16 AM:
{quote}
i've added extra abort() statement to safeguard against cases where the caller
does not handle this exception explicitly.
{quote}
Then this is the caller's fault? Why not fix it at the caller side?
Or we could remove the abort regionserver semantics of
DroppedSnapshotException so the caller do not need to deal with it anymore?
We'd better make the rule clear otherwise people may get confusing...
was (Author: apache9):
{code}
i've added extra abort() statement to safeguard against cases where the caller
does not handle this exception explicitly.
{code}
Then this is the caller's fault? Why not fix it at the caller side?
Or we could remove the abort regionserver semantics of
DroppedSnapshotException so the caller do not need to deal with it anymore?
We'd better make the rule clear otherwise people may get confusing...
Interrupt to flush from TableFlushProcedure causes dataloss in ITBLL
Key: HBASE-13877
URL: https://issues.apache.org/jira/browse/HBASE-13877
Project: HBase
Issue Type: Bug
Reporter: Enis Soztutar
Assignee: Enis Soztutar
Priority: Blocker
Fix For: 2.0.0, 1.2.0, 1.1.1
Attachments: hbase-13877_v1.patch, hbase-13877_v2-branch-1.1.patch
ITBLL with 1.25B rows failed for me (and Stack as reported in
https://issues.apache.org/jira/browse/HBASE-13811?focusedCommentId=14577834page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-14577834)
HBASE-13811 and HBASE-13853 fixed an issue with WAL edit filtering.
The root cause this time seems to be different. It is due to procedure based
flush interrupting the flush request in case the procedure is cancelled from
an exception elsewhere. This leaves the memstore snapshot intact without
aborting the server. The next flush, then flushes the previous memstore with
the current seqId (as opposed to seqId from the memstore snapshot). This
creates an hfile with larger seqId than what its contents are. Previous
behavior in 0.98 and 1.0 (I believe) is that after flush prepare and
interruption / exception will cause RS abort.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
