Fix db_stress failure in async_io in FilePrefetchBuffer #10949
Conversation
akankshamahajan15
force-pushed
the
async_io
branch
from
02a65de
to
80ecd81
Compare
akankshamahajan15
changed the title
|
@akankshamahajan15 has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator. |
akankshamahajan15
force-pushed
the
async_io
branch
from
80ecd81
to
7797967
Compare
|
@akankshamahajan15 has updated the pull request. You must reimport the pull request before landing. |
|
@akankshamahajan15 has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator. |
|
I am not able to figure out what causes it to call |
Summary: Test Plan: Reviewers: Subscribers: Tasks: Tags:
akankshamahajan15
force-pushed
the
async_io
branch
from
7797967
to
8fa53aa
Compare
|
@akankshamahajan15 has updated the pull request. You must reimport the pull request before landing. |
|
@akankshamahajan15 has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator. |
|
@akankshamahajan15 Can you explain the problem a bit more in the description? What was the sequence of events that caused the db_stress failure? |
Oh interesting. Its possible a read by another thread might have loaded the block into the cache in the meantime. In that case, will the reseek update prev_len_ and prev_offset_ but explicit_prefetch_submitted_ will remain true? |
Oh. I didn't know another thread could load the block but that seems like the only case. Yes we don't set explicit_prefetch_submitted_ = false in that case. Let me set it to false and see if that works. |
Summary: Test Plan: Reviewers: Subscribers: Tasks: Tags:
|
@akankshamahajan15 has updated the pull request. You must reimport the pull request before landing. |
|
@akankshamahajan15 has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator. |
|
@akankshamahajan15 has updated the pull request. You must reimport the pull request before landing. |
|
@akankshamahajan15 has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator. |
Summary: Fix db_stress failure in async_io in FilePrefetchBuffer. From the logs, assertion was caused when - prev_offset_ = offset but somehow prev_len != 0 and explicit_prefetch_submitted_ = true. That scenario is when we send async request to prefetch buffer during seek but in second seek that data is found in cache. prev_offset_ and prev_len_ get updated but we were not setting explicit_prefetch_submitted_ = false because of which buffers were getting out of sync. It's possible a read by another thread might have loaded the block into the cache in the meantime. Particular assertion example: ``` prev_offset: 0, prev_len_: 8097 , offset: 0, length: 8097, actual_length: 8097 , actual_offset: 0 , curr_: 0, bufs_[curr_].offset_: 4096 ,bufs_[curr_].CurrentSize(): 48541 , async_len_to_read: 278528, bufs_[curr_].async_in_progress_: false second: 1, bufs_[second].offset_: 282624 ,bufs_[second].CurrentSize(): 0, async_len_to_read: 262144 ,bufs_[second].async_in_progress_: true , explicit_prefetch_submitted_: true , copy_to_third_buffer: false ``` As we can see curr_ was expected to read 278528 but it read 48541. Also buffers are out of sync. Also `explicit_prefetch_submitted_` is set true but prev_len not 0. Pull Request resolved: #10949 Test Plan: - Ran db_bench for regression to make sure there is no regression; - Ran db_stress failing without this fix, - Ran build-linux-mini-crashtest 7- 8 times locally + CircleCI Reviewed By: anand1976 Differential Revision: D41257786 Pulled By: akankshamahajan15 fbshipit-source-id: 1d100f94f8c06bbbe4cc76ca27f1bbc820c2494f
Summary: Fix db_stress failure in async_io in FilePrefetchBuffer. From the logs, assertion was caused when - prev_offset_ = offset but somehow prev_len != 0 and explicit_prefetch_submitted_ = true. That scenario is when we send async request to prefetch buffer during seek but in second seek that data is found in cache. prev_offset_ and prev_len_ get updated but we were not setting explicit_prefetch_submitted_ = false because of which buffers were getting out of sync. It's possible a read by another thread might have loaded the block into the cache in the meantime. Particular assertion example: ``` prev_offset: 0, prev_len_: 8097 , offset: 0, length: 8097, actual_length: 8097 , actual_offset: 0 , curr_: 0, bufs_[curr_].offset_: 4096 ,bufs_[curr_].CurrentSize(): 48541 , async_len_to_read: 278528, bufs_[curr_].async_in_progress_: false second: 1, bufs_[second].offset_: 282624 ,bufs_[second].CurrentSize(): 0, async_len_to_read: 262144 ,bufs_[second].async_in_progress_: true , explicit_prefetch_submitted_: true , copy_to_third_buffer: false ``` As we can see curr_ was expected to read 278528 but it read 48541. Also buffers are out of sync. Also `explicit_prefetch_submitted_` is set true but prev_len not 0. Pull Request resolved: #10949 Test Plan: - Ran db_bench for regression to make sure there is no regression; - Ran db_stress failing without this fix, - Ran build-linux-mini-crashtest 7- 8 times locally + CircleCI Reviewed By: anand1976 Differential Revision: D41257786 Pulled By: akankshamahajan15 fbshipit-source-id: 1d100f94f8c06bbbe4cc76ca27f1bbc820c2494f
Summary: Fix db_stress failure in async_io in FilePrefetchBuffer. From the logs, assertion was caused when - prev_offset_ = offset but somehow prev_len != 0 and explicit_prefetch_submitted_ = true. That scenario is when we send async request to prefetch buffer during seek but in second seek that data is found in cache. prev_offset_ and prev_len_ get updated but we were not setting explicit_prefetch_submitted_ = false because of which buffers were getting out of sync. It's possible a read by another thread might have loaded the block into the cache in the meantime. Particular assertion example: ``` prev_offset: 0, prev_len_: 8097 , offset: 0, length: 8097, actual_length: 8097 , actual_offset: 0 , curr_: 0, bufs_[curr_].offset_: 4096 ,bufs_[curr_].CurrentSize(): 48541 , async_len_to_read: 278528, bufs_[curr_].async_in_progress_: false second: 1, bufs_[second].offset_: 282624 ,bufs_[second].CurrentSize(): 0, async_len_to_read: 262144 ,bufs_[second].async_in_progress_: true , explicit_prefetch_submitted_: true , copy_to_third_buffer: false ``` As we can see curr_ was expected to read 278528 but it read 48541. Also buffers are out of sync. Also `explicit_prefetch_submitted_` is set true but prev_len not 0. Pull Request resolved: facebook/rocksdb#10949 Test Plan: - Ran db_bench for regression to make sure there is no regression; - Ran db_stress failing without this fix, - Ran build-linux-mini-crashtest 7- 8 times locally + CircleCI Reviewed By: anand1976 Differential Revision: D41257786 Pulled By: akankshamahajan15 fbshipit-source-id: 1d100f94f8c06bbbe4cc76ca27f1bbc820c2494f
Summary: Fix db_stress failure in async_io in FilePrefetchBuffer. From the logs, assertion was caused when - prev_offset_ = offset but somehow prev_len != 0 and explicit_prefetch_submitted_ = true. That scenario is when we send async request to prefetch buffer during seek but in second seek that data is found in cache. prev_offset_ and prev_len_ get updated but we were not setting explicit_prefetch_submitted_ = false because of which buffers were getting out of sync. It's possible a read by another thread might have loaded the block into the cache in the meantime. Particular assertion example: ``` prev_offset: 0, prev_len_: 8097 , offset: 0, length: 8097, actual_length: 8097 , actual_offset: 0 , curr_: 0, bufs_[curr_].offset_: 4096 ,bufs_[curr_].CurrentSize(): 48541 , async_len_to_read: 278528, bufs_[curr_].async_in_progress_: false second: 1, bufs_[second].offset_: 282624 ,bufs_[second].CurrentSize(): 0, async_len_to_read: 262144 ,bufs_[second].async_in_progress_: true , explicit_prefetch_submitted_: true , copy_to_third_buffer: false ``` As we can see curr_ was expected to read 278528 but it read 48541. Also buffers are out of sync. Also `explicit_prefetch_submitted_` is set true but prev_len not 0. Pull Request resolved: facebook/rocksdb#10949 Test Plan: - Ran db_bench for regression to make sure there is no regression; - Ran db_stress failing without this fix, - Ran build-linux-mini-crashtest 7- 8 times locally + CircleCI Reviewed By: anand1976 Differential Revision: D41257786 Pulled By: akankshamahajan15 fbshipit-source-id: 1d100f94f8c06bbbe4cc76ca27f1bbc820c2494f
Summary: Fix db_stress failure in async_io in FilePrefetchBuffer. From the logs, assertion was caused when - prev_offset_ = offset but somehow prev_len != 0 and explicit_prefetch_submitted_ = true. That scenario is when we send async request to prefetch buffer during seek but in second seek that data is found in cache. prev_offset_ and prev_len_ get updated but we were not setting explicit_prefetch_submitted_ = false because of which buffers were getting out of sync. It's possible a read by another thread might have loaded the block into the cache in the meantime. Particular assertion example: ``` prev_offset: 0, prev_len_: 8097 , offset: 0, length: 8097, actual_length: 8097 , actual_offset: 0 , curr_: 0, bufs_[curr_].offset_: 4096 ,bufs_[curr_].CurrentSize(): 48541 , async_len_to_read: 278528, bufs_[curr_].async_in_progress_: false second: 1, bufs_[second].offset_: 282624 ,bufs_[second].CurrentSize(): 0, async_len_to_read: 262144 ,bufs_[second].async_in_progress_: true , explicit_prefetch_submitted_: true , copy_to_third_buffer: false ``` As we can see curr_ was expected to read 278528 but it read 48541. Also buffers are out of sync. Also `explicit_prefetch_submitted_` is set true but prev_len not 0. Pull Request resolved: facebook/rocksdb#10949 Test Plan: - Ran db_bench for regression to make sure there is no regression; - Ran db_stress failing without this fix, - Ran build-linux-mini-crashtest 7- 8 times locally + CircleCI Reviewed By: anand1976 Differential Revision: D41257786 Pulled By: akankshamahajan15 fbshipit-source-id: 1d100f94f8c06bbbe4cc76ca27f1bbc820c2494f
Summary: Fix db_stress failure in async_io in FilePrefetchBuffer. From the logs, assertion was caused when - prev_offset_ = offset but somehow prev_len != 0 and explicit_prefetch_submitted_ = true. That scenario is when we send async request to prefetch buffer during seek but in second seek that data is found in cache. prev_offset_ and prev_len_ get updated but we were not setting explicit_prefetch_submitted_ = false because of which buffers were getting out of sync. It's possible a read by another thread might have loaded the block into the cache in the meantime. Particular assertion example: ``` prev_offset: 0, prev_len_: 8097 , offset: 0, length: 8097, actual_length: 8097 , actual_offset: 0 , curr_: 0, bufs_[curr_].offset_: 4096 ,bufs_[curr_].CurrentSize(): 48541 , async_len_to_read: 278528, bufs_[curr_].async_in_progress_: false second: 1, bufs_[second].offset_: 282624 ,bufs_[second].CurrentSize(): 0, async_len_to_read: 262144 ,bufs_[second].async_in_progress_: true , explicit_prefetch_submitted_: true , copy_to_third_buffer: false ``` As we can see curr_ was expected to read 278528 but it read 48541. Also buffers are out of sync. Also `explicit_prefetch_submitted_` is set true but prev_len not 0. Pull Request resolved: facebook/rocksdb#10949 Test Plan: - Ran db_bench for regression to make sure there is no regression; - Ran db_stress failing without this fix, - Ran build-linux-mini-crashtest 7- 8 times locally + CircleCI Reviewed By: anand1976 Differential Revision: D41257786 Pulled By: akankshamahajan15 fbshipit-source-id: 1d100f94f8c06bbbe4cc76ca27f1bbc820c2494f
Summary: Fix db_stress failure in async_io in FilePrefetchBuffer.
From the logs, assertion was caused when
It's possible a read by another thread might have loaded the block into the cache in the meantime.
Particular assertion example:
As we can see curr_ was expected to read 278528 but it read 48541. Also buffers are out of sync.
Also
explicit_prefetch_submitted_is set true but prev_len not 0.Test Plan:
Reviewers:
Subscribers:
Tasks:
Tags: