| 1 |
>> > Has anyone else noticed this sort of behavior from apache-2.2.21-r1: |
| 2 |
>> > |
| 3 |
>> > # /etc/init.d/apache2 restart |
| 4 |
>> > |
| 5 |
>> > * Stopping apache2 ... [ ok ] |
| 6 |
>> > * Starting apache2 ... |
| 7 |
>> > * start-stop-daemon: /usr/sbin/apache2 is already running [ ok ] |
| 8 |
>> > |
| 9 |
>> > # /etc/init.d/apache2 restart |
| 10 |
>> > |
| 11 |
>> > * apache2 not running (no pid file) |
| 12 |
>> > * Starting apache2 ... [ ok ] |
| 13 |
>> > |
| 14 |
>> > It probably happens about half the time I restart apache2. Nothing in |
| 15 |
>> > /var/log/apache2/error_log besides: |
| 16 |
>> > |
| 17 |
>> > [notice] caught SIGTERM, shutting down |
| 18 |
>> > [notice] Apache/2.2.21 (Unix) configured -- resuming normal operations |
| 19 |
>> > |
| 20 |
>> > - Grant |
| 21 |
>> |
| 22 |
>> I've seen that for years. I always assumed it was because the apache2 |
| 23 |
>> stop didn't complete before the apache2 start tried to start when I use |
| 24 |
>> /etc/init.d/apache2 restart. |
| 25 |
>> |
| 26 |
>> Instead I do an /etc/init.d/apache2 stop; ps aux | egrep apache until I |
| 27 |
>> see all the threads are gone and then a /etc/init.d/apache2 start |
| 28 |
> |
| 29 |
> +1 |
| 30 |
> |
| 31 |
> If there aren't many threads a quick restart will work fine, but if there are |
| 32 |
> many threads hanging around then you may be better off to stop, wait, then |
| 33 |
> start. |
| 34 |
|
| 35 |
Thanks fellas, that makes perfect sense. Would it make sense for me |
| 36 |
to request a change to the initscript that waits until all threads |
| 37 |
have stopped before starting during a restart? |
| 38 |
|
| 39 |
- Grant |
Archives