On 2009-01-20, Collins, Kevin [BEELINE] <[email protected]> wrote:
>
> So my assumption would be that when the 1-minute laod average is less
> than max_load, it would start accepting connections again.
>

Yes, of course, that's what I assumed too.. until I tried to verify it
by setting max_load     = 1, trigger it by running "top -d 0" a little while,
stop the 'top' process to let the load go and notice that after waiting
at least 10 minutes, it's still not re-enabled..

..waited more than an hour, and still the service is disabled.

Tried "service xinetd reload" -- still not re-enabled.

But after a full "service xinetd restart" it starts working again.

Full log:

Jan 20 18:18:46 minimac xinetd[26428]: Started working: 1 available service
Jan 20 18:19:49 minimac xinetd[26428]: START: tftp pid=26463 from=127.0.0.1
Jan 20 18:21:08 minimac xinetd[26428]: EXIT: tftp signal=15 pid=26463 
duration=79(sec)
Jan 20 18:21:12 minimac xinetd[26428]: refused connect from 127.0.0.1 due to 
excessive load
Jan 20 18:21:12 minimac xinetd[26428]: FAIL: tftp load from=127.0.0.1

<try lots of tftp-connections, no respons>

Jan 20 19:32:33 minimac xinetd[26428]: Starting reconfiguration
Jan 20 19:32:33 minimac xinetd[26428]: Swapping defaults
Jan 20 19:32:33 minimac xinetd[26428]: readjusting service tftp
Jan 20 19:32:33 minimac xinetd[26428]: Reconfigured: new=0 old=1 dropped=0 
(services)

<try lots of tftp-connections, no respons>

Jan 20 19:33:21 minimac xinetd[26428]: Exiting...
Jan 20 19:33:21 minimac xinetd[26653]: xinetd Version 2.3.14 started with 
libwrap loadavg labeled-networking options compiled in.
Jan 20 19:33:21 minimac xinetd[26653]: Started working: 1 available service
Jan 20 19:33:26 minimac xinetd[26653]: START: tftp pid=26656 from=127.0.0.1
Jan 20 19:33:26 minimac xinetd[26653]: EXIT: tftp status=0 pid=26656 
duration=0(sec)

<try tftp-connection, and now it works>

BTW: I'm running the tftpd server with argument "-t 0" to force it to
exit after every connection. Otherwise the max_load has no effect.

This test was on fedora-9, but I did the same exercise on RHEL5u2 earlier today.
I'd very much like to post a bug report, except I can't find it documented that
it should resume the service after the load goes down again.. :-(


  -jf

_______________________________________________
rhelv5-list mailing list
[email protected]
https://www.redhat.com/mailman/listinfo/rhelv5-list

Reply via email to