Bug #57914 | invalid sleep values results in different error messages in windows and linux | ||
---|---|---|---|
Submitted: | 2 Nov 2010 6:57 | Modified: | 2 Dec 2010 0:09 |
Reporter: | Satya B | Email Updates: | |
Status: | Closed | Impact on me: | |
Category: | MySQL Enterprise Backup | Severity: | S3 (Non-critical) |
Version: | 3.5.2 | OS: | Any |
Assigned to: | Satya B | CPU Architecture: | Any |
Tags: | innobackup, MEB 3.5.2, mysqlbackup |
[2 Nov 2010 6:57]
Satya B
[8 Nov 2010 10:56]
Satya B
Suggested patch is: === modified file 'meb-innobackup-c/src/innobackup.c' --- meb-innobackup-c/src/innobackup.c 2010-10-19 11:45:13 +0000 +++ meb-innobackup-c/src/innobackup.c 2010-11-03 10:27:48 +0000 @@ -3311,6 +3311,7 @@ { /* Process state has been changed */ case WAIT_OBJECT_0: + print_message(NO_TIMESTAMP, ERROR, "ibbackup process has died"); goto err20; /* Wait Failed */ case WAIT_FAILED: After applying the suggested patch, innobackup_logs_field will now fail in windows due to a difference in error code. SELECT exit_state, last_error, last_error_code FROM mysql.backup_history; exit_state last_error last_error_code -FAILURE ibbackup process has died 2 +FAILURE ibbackup process has died 13 The last_error_code is generated by errno. #define ENOENT 2 /* No such file or directory */ #define EACCES 13 /* Permission denied */ The errno is 2 in linux and 13 in windows. The errno 2 value may not be correct as the mysqlbackup process dies because the child process (ibbackup) cannot execute ( due to an invalid sleep value). On the other hand, 13 looks appropriate as child wasn't executed properly. I wonder how errno is mapped in windows as errno is in UNIX specification only and value of 13 in windows is surprising. Using windows specific api, GetLastError() gives 0. This may be because of the fact the error was thrown by child and not by the parent mysqlbackup process. Q: == 1. This field is only for informational purpose. So it is Okay to Mask this in testcase. 2. Windows specific api GetLastError() returns 0 which is worse than what errno displays. 3. Any other solutions to this?
[11 Nov 2010 15:39]
Ingo Strüwing
Approved. No comments. Test case fix (suppression of error number) required.
[18 Nov 2010 6:21]
Satya B
branch: meb-3.5.next revno: 252 revision-id: satya.bn@sun.com-20101109073201-a1lj8j1vbu8b0593