| Bug #30021 | Server init.d script stop leaves dead server process running | ||
|---|---|---|---|
| Submitted: | 25 Jul 2007 1:06 | Modified: | 31 Aug 2007 3:13 |
| Reporter: | Adam Donnison | Email Updates: | |
| Status: | No Feedback | Impact on me: | |
| Category: | MySQL Server | Severity: | S3 (Non-critical) |
| Version: | 5.1.20 | OS: | Linux (Fedora Core) |
| Assigned to: | CPU Architecture: | Any | |
[31 Jul 2007 3:13]
Valeriy Kravchuk
Thank you for a problem report. Please, send the results of: uname -a cat /etc/issue getconf GNU_LIBPTHREAD_VERSION getconf GNU_LIBC_VERSION from all the system affected and NOT affected by this bug.
[31 Aug 2007 23:00]
Bugs System
No feedback was provided for this bug for over a month, so it is being suspended automatically. If you are able to provide the information that was originally requested, please do so and change the status of the bug back to "Open".
Description: On 4 out of 6 servers running 5.1.20, 'service mysql stop' or running '/etc/init.d/mysql stop' leaves both mysqld and mysqld_safe processes. Starting the service again results in multiple copies of mysqld and mysqld_safe running. kill with default signal (SIGTERM) does nothing, SIGKILL is required. Interestingly running strace against the mysqld process also causes it to die with the following output: Process 3594 attached - interrupt to quit select(16, [14 15], NULL, NULL, NULL) = -1 EBADF (Bad file descriptor) gettimeofday({1185316956, 508069}, NULL) = 0 exit_group(0) = ? Process 3594 detached How to repeat: As root on a Fedora Core system with mysql running, issue the command: service mysql stop Wait until you get the OK, then issue the command: ps -ef|grep mysqld Shows both mysqld and mysqld_safe running. Suggested fix: Add a valid timeout to the select call on the unix and TCP sockets. It may be that the infinite timeout is causing the exit to be blocked until some other process causes an interrupt in the select.