-
Type:
Bug
-
Resolution: Duplicate
-
Priority:
Major - P3
-
None
-
Affects Version/s: None
-
Component/s: None
-
None
-
3
-
Storage - Tora 2020-05-04
File corruption in the history store can result in a wiredtiger panic in checkpoint, furthermore it currently won't suggest salvaging the file. Ideally we should detect and return a nicer error when finding corruption in the metadata. Below is output from the wt utility when we detect history store corruption:
[1583111051:484932][35317:0x7fecb9a15740], wt, file:WiredTigerHS.wt, WT_SESSION.checkpoint: __posix_file_read, 421: RESTART/WiredTigerHS.wt: handle-read: pread: failed to read 4096 bytes at offset 0: WT_ERROR: non-specific WiredTiger error [1583111051:484986][35317:0x7fecb9a15740], wt, file:test_txn19_meta_0.wt, WT_SESSION.checkpoint: __wt_block_checkpoint_resolve, 862: test_txn19_meta_0.wt: the checkpoint failed, the system must restart: Invalid argument [1583111051:485001][35317:0x7fecb9a15740], wt, file:test_txn19_meta_0.wt, WT_SESSION.checkpoint: __wt_panic, 490: the process must exit and restart: WT_PANIC: WiredTiger library panic [1583111051:485008][35317:0x7fecb9a15740], wt, file:test_txn19_meta_0.wt, WT_SESSION.checkpoint: __wt_abort, 28: aborting WiredTiger library