1 |
On 21/02/2014 09:03, Yuri K. Shatroff wrote: |
2 |
>> Your idea instantly fails as the rc-service author has no idea of what |
3 |
>> you defined ${SERVICE} to be and no way to determine what it is now. |
4 |
> |
5 |
> Yes, the rc-service author does not have any idea because he is not |
6 |
> requested to. |
7 |
> ${SERVICE} obviously comes from `rc-service status ${SERVICE}` . |
8 |
> The result (e.g. tail -n {$LINES} ${SERVICE}.log) is achieved by: |
9 |
> 1. putting LINES= in /etc/conf.d/${SERVICE} |
10 |
> 2. setting up ${SERVICE}.log with syslog. (or putting LOGFILE=... and |
11 |
> doing `tail -n ${LINES} ${LOGFILE}, or even LAST_LOG_CMD=`mysql -qe |
12 |
> 'SELECT ... FROM log.log ORDER BY date DESC LIMIT ${LINES}'`, or |
13 |
> *whatever*) |
14 |
> 3. adding this `tail -n ...` or whatever call to the init script . |
15 |
> 4. voila. |
16 |
> |
17 |
> If you feel I'm again entirely wrong please point out why. |
18 |
|
19 |
|
20 |
The faults with your comments are many, and I'm not going to detail them |
21 |
as that's not my job. I'm going to let you figure it out for yourself in |
22 |
production why your entire approach is wrong, and simply leave you with |
23 |
this: |
24 |
|
25 |
You violate DRY. |
26 |
|
27 |
You expect the sysadmin to know they must make changes in a restart |
28 |
config file when they tweak the syslogger so that somehow the init |
29 |
script continues to get it right. Trust me, sysadmins are not going to |
30 |
remember to do that, because expecting them to is off the wall crazy. |
31 |
|
32 |
I repeat what I and Canek said earlier: |
33 |
|
34 |
You've never actually DONE any of this in real life, right? |
35 |
|
36 |
|
37 |
-- |
38 |
Alan McKinnon |
39 |
alan.mckinnon@×××××.com |