Major - P3 alexander.gorrod reported this: There was another failure with a slightly different signature:
Beginning cycle 5/5 Suspend child Check DB [1537152626:197438][31983:0x403b300], file:A306-0.wt, txn-recover: __posix_open_file, 715: WT_TEST.random-directio.CHECK/A306-0.wt: handle-open: open: No such file or directory [1537152626:222695][31983:0x403b300], file:A306-0.wt, txn-recover: __txn_op_apply, 280: operation apply failed during recovery: operation type 4 at LSN 1/1324544: No such file or directory [1537152626:225091][31983:0x403b300], file:A306-0.wt, txn-recover: __wt_txn_recover, 723: Recovery failed: No such file or directory Copy database home directory using direct I/O to run recovery,
This was originally reported in WT-4322. I think this is not related to the 'dd copy' problem we talk about WT-4322 because that problem appears to be related to the version of dd. The valgrind failure here happened on margay, with dd version 8.22. Also, using a pull request with the test rewritten to not rely on dd, the failure still occurs, and I have directories that seem to show that files were copied correctly. Here's another failure:
Beginning cycle 2/5 Suspend child Check DB [1537290574:465113][25422:0x403b300], file:A202-2.wt, txn-recover: __posix_open_ file, 715: WT_TEST.random-directio.CHECK/A202-2.wt: handle-open: open: No such f ile or directory [1537290574:489113][25422:0x403b300], file:A202-2.wt, txn-recover: __txn_op_appl y, 280: operation apply failed during recovery: operation type 4 at LSN 1/104012 8: No such file or directory [1537290574:491277][25422:0x403b300], file:A202-2.wt, txn-recover: __wt_txn_reco ver, 723: Recovery failed: No such file or directory
I only got this to occur in the context of make check, running the commands used for the valgrind job, so I'm not sure if the test is running under valgrind as well.