227 Commits

Author SHA1 Message Date
Meng Xu
2febbe74ce FastRestoreApplier:Fix conflict range inverted due to invalid memory access
Rerpot error on loader and applier if not error_code_operation_cancelled error ever happens
2020-09-09 14:40:21 -07:00
Meng Xu
e69ee87f33 Refactor blob failure to simulateBlobFailure function 2020-09-09 10:54:21 -07:00
Meng Xu
5880bd9fa7 BackupContainer:Fix getSnapshotFileKeyRange_impl on blob http error 2020-09-09 08:47:43 -07:00
Meng Xu
83d1350d8d FastRestore:Handle retriable blob error 2020-09-09 07:29:17 -07:00
Meng Xu
6d58064964 Buggify _parsePartitionedLogFileOnLoader with blob http error 2020-09-09 06:46:08 -07:00
Meng Xu
f3986d201d Buggify backup decodeRangeFileBlock with blob connection error 2020-09-09 06:40:41 -07:00
Meng Xu
7a29a3157f FastRestore:Fix various bugs discovered by enhanced simulation
1. sendMutation request can be dispatched when the version batch has finished and its data has been deleted;
2. Request scheduler on loader may get into infinite loop because FASTRESTORE_SCHED_LOAD_REQ_BATCHSIZE knob can be set to 0 in simulation
2020-08-26 15:53:18 -07:00
Meng Xu
f1bd2a18ed Resovle review comments: No functional change 2020-08-26 10:30:31 -07:00
Meng Xu
bd7c07436b FastRestore:Add batchIndex to RestoreAsset for better performance tracking 2020-08-25 09:34:18 -07:00
Meng Xu
778daf20c0 FastRestore:Fix incorrect assert 2020-08-24 19:59:56 -07:00
Meng Xu
996ba2374c FastREstore:Fix:Incorrect condition in printing out FastRestoreLoaderSendMutationToApplierDoneTooLate 2020-08-24 17:27:40 -07:00
Meng Xu
e94261efe5 FastRestore:LoaderScheduler:Add validation on too old requests 2020-08-24 16:32:16 -07:00
Meng Xu
6e3e36c8fc FastRestore:RequestScheduler:Minor code style improvement 2020-08-24 10:45:46 -07:00
Meng Xu
d9ea14ea6c FastRestore:fix:loader can receive reply from vb that has been processed and deleted 2020-08-19 10:39:49 -07:00
Meng Xu
88b94a3e59 FastRestore:Fix:Delayed request can be skipped from processing
When a request of sending mutations to appliers is delayed and the request for future version batches have been processed,
loader will increase its counter. The previous implementation will use the counter to decide if a request is too old and
have been processed. If the counter is larger than the batch index of the request, the request will be ignored although
it may not have been processed.

This causes correctness test failure for fast restore.
2020-08-18 23:48:06 -07:00
Meng Xu
e1d150662f FastRestore:small fix on compile error 2020-08-18 16:15:11 -07:00
Meng Xu
a32cf7d293 FastRestore:Loader:Add sanity check before remove batch data 2020-08-18 15:44:22 -07:00
Meng Xu
9b2f667bbe FastRestore:Fix uninitialized variable 2020-08-18 11:58:57 -07:00
Meng Xu
4a0315483b FastRestore:Safeguard when request of earlier vb may be sent after the vb has finished 2020-08-17 22:20:54 -07:00
Meng Xu
7b7490efe7 FastRestore:Debug trace for seg fault 2020-08-17 20:34:33 -07:00
Meng Xu
01255b7ead FastRestoreLoader:Sched:Add trace field and misc code style improvement 2020-08-16 09:43:04 -07:00
Meng Xu
be75fd4894 Workaround flow limitation:state variable must be initialized at the beginning of actors 2020-08-16 08:45:13 -07:00
Meng Xu
c2dfc5a00a Fix compile error 2020-08-16 08:13:28 -07:00
Meng Xu
cb0e44ff42 FastRestoreLoader:Sched send mutations per loadingParam requests 2020-08-16 07:38:51 -07:00
Meng Xu
98167a6d20 FastRestoreLoader:Fix:sched may stuck due to race condition 2020-08-16 06:29:09 -07:00
Meng Xu
015397be74 FastRestore:LoaderScheduler:Control number of load request to release at once 2020-08-15 22:18:48 -07:00
Meng Xu
1e571a5a1a FastRestore:Loader:Kick off scheduler when loader starts to have new requests 2020-08-15 21:57:00 -07:00
Meng Xu
a035e7a872 FastRestore:LoaderScheduler:Add knobs 2020-08-15 21:44:51 -07:00
Meng Xu
7e302b5910 FastRestore:Loader:Add scheduler to dispatch requests based on resource utilization and request priority 2020-08-15 21:21:51 -07:00
Meng Xu
c27542d4e9 FR:SampleBackup:Send reply when receive duplicate msg 2020-08-07 16:50:36 -07:00
Meng Xu
fccebd9b80 FR:Use ByteSampleInfo to accurately sample backup 2020-08-06 20:45:46 -07:00
Meng Xu
87bde08695 FR:fix:sending samples may have broken promise error
The error should not skip the ack reply from loader to controller, saying
the file has been loaded.
Otherwise, restore controller can get stuck at waiting for the ack reply.
2020-08-06 10:41:18 -07:00
Meng Xu
859c3145e2 FR:Fix samples not send reply and false positive failures 2020-08-05 11:32:04 -07:00
Meng Xu
879998d01b FR:fix compilation error 2020-08-04 23:04:38 -07:00
Meng Xu
7992cef025 FR:Fix sample network pkg can be too big 2020-08-04 22:35:21 -07:00
Meng Xu
fec3de9b7b Fix RestoreLoader loadedLogBytes counter 2020-07-31 16:00:15 -07:00
Jingyu Zhou
7279beb662 Add old splitMutation for unit test 2020-07-08 17:57:20 -07:00
Jingyu Zhou
d6dd88c2b9 Move building of KeyRangeMap out of the loop
So that it is only done once.
2020-07-08 14:16:10 -07:00
Jingyu Zhou
a187eeebee Add assertions back and comments for splitMutation 2020-07-08 11:08:23 -07:00
Jingyu Zhou
e78cc9ee4b Use KeyRangeMap for splitMutation
This simplifies code by reuse KeyRangeMap class.
2020-07-08 09:53:34 -07:00
Meng Xu
a5febfbc75 Remove spurious comments 2020-06-29 10:26:01 -07:00
Meng Xu
8f5ee1f4aa Refine trace events for better perf diagnosis 2020-06-29 10:22:07 -07:00
Meng Xu
82dfb5ce3f FastRestore:Update process metrics for restore master 2020-06-28 12:37:04 -07:00
Meng Xu
bc98c84346 RestoreLoader release data early and revert Lower priority for RestoreApplierReceiveMutations actor
A quick evalution shows lowering priority for receive mutation actor does not help restore speed but hurt it.
2020-06-28 11:12:06 -07:00
Meng Xu
026a735703 Test perf without gracefully clear memory 2020-06-27 16:15:56 -07:00
Meng Xu
d6381de16f RestoreLoader:Clear memory per loading param 2020-06-27 15:51:19 -07:00
Meng Xu
f18a0a242d FastRestore:Loader frees mutation memory immediately after send mutations 2020-06-27 15:19:08 -07:00
Meng Xu
e57dba00bd FastRestore:Lower priority for RestoreApplierReceiveMutations actor 2020-06-27 15:16:38 -07:00
Meng Xu
5860a5b4db FastRestore:Suppress or mute spammy trace events 2020-06-24 22:10:54 -07:00
Meng Xu
c13ba7aeb4 FastRestore:Cosmetic code style improvement
Also mute some trace events
2020-06-24 21:25:37 -07:00