[ntp:questions] Ntpd in uninterruptible sleep?
A C
agcarver+ntp at acarver.net
Wed Oct 26 06:41:22 UTC 2011
On 10/25/2011 20:46, Steve Kostecke wrote:
> On 2011-10-25, A C<agcarver+ntp at acarver.net> wrote:
>> On 10/25/2011 14:53, unruh wrote:
>>>
>>> ?? So why are you using X at all? Xterm can surely be replaced by the
>>> console. And xclock by running date to tell you what time the machine
>>> thinks it is (I assume you use that to check to see if ntpd has stopped
>>> controlling the clock)
>>
>> A few reasons for not using console:
>> 1. I don't have a monitor to devote to the machine that works at
>> standard Sun framebuffer refresh rates. With the odd resolution and
>> rate of the CG6 most inexpensive LCDs just fall over and refuse to sync.
>> My one monitor that can sync is also the nice one that I use for my
>> regular desktop machine.
>>
>> 2. Even with a monitor on the machine there's a bug in the console
>> driver in NetBSD that locks up the console getty if redraws occur too
>> quickly. It's a known bug on the sparc hardware but hard to track down.
>> It doesn't affect anything else except the getty on the console.
>>
>> 3. I can leave the Xterm on my screen on my everyday machine and be able
>> to use it when I need to without needing to reach for another keyboard.
>> Plus anything I do on the machine, especially during debugging right
>> now, I can copy and paste to post to the list or save for later.
>
> Wouldn't it be easier to just open an Xterm on your everyday machine and
> ssh into the time server?
>
That's almost what I'm doing. I just happen to have a quick icon on my
regular machine that creates an ssh session and sends back an xterm.
It's really not causing much if any overhead on the Sun machine. Here's
the output from top with top itself running inside the xterm so
refreshes are occuring and causing the xterm and ssh to use a bit of
CPU. There isn't much load at all, mostly the garbage collector. I
haven't had a chance yet to put the memory back in the system so it's
still running with 16 MB.
> load averages: 0.17, 0.06, 0.01; up 16+12:49:20 06:38:54
> 22 processes: 21 sleeping, 1 on CPU
> CPU states: 11.4% user, 0.0% nice, 7.3% system, 4.9% interrupt, 76.3% idle
> Memory: 4824K Act, 2440K Inact, 300K Wired, 2180K Exec, 1412K File, 460K Free
> Swap: 128M Total, 8820K Used, 120M Free
>
> PID USERNAME PRI NICE SIZE RES STATE TIME WCPU CPU COMMAND
> 0 root 125 0 0K 5392K cachegc 158:13 3.56% 3.56% [system]
> 29574 root 85 0 5416K 896K pause 1:54 1.76% 1.76% ntpd
> 408 agcarver 85 0 11M 752K select 2:40 1.56% 1.56% sshd
> 1047 root 85 0 3448K 496K wait 0:23 0.98% 0.98% bash
> 533 agcarver 85 0 11M 668K select 118:57 0.78% 0.78% sshd
> 323 nobody 90 -10 13M 576K select 41.9H 0.00% 0.00% gpsd
> 415 agcarver 85 0 7172K 504K select 105:06 0.00% 0.00% xclock
> 410 agcarver 85 0 7328K 836K select 31:01 0.00% 0.00% xterm
> 335 root 85 0 5708K 260K kqueue 3:37 0.00% 0.00% master
> 391 postfix 85 0 5720K 308K kqueue 2:39 0.00% 0.00% qmgr
> 400 root 85 0 3372K 244K nanoslp 1:43 0.00% 0.00% cron
> 115 root 85 0 3468K 288K kqueue 1:23 0.00% 0.00% syslogd
> 508 root 83 0 11M 0K netio 0:06 0.00% 0.00% <sshd>
> 363 root 85 0 11M 0K netio 0:05 0.00% 0.00% <sshd>
> 34 root 85 0 3292K 0K pause 0:01 0.00% 0.00% <csh>
> 926 root 43 0 3624K 1112K CPU 0:00 0.00% 0.00% top
> 28716 postfix 85 0 5708K 616K kqueue 0:00 0.00% 0.00% pickup
> 1 root 85 0 3472K 60K wait 0:00 0.00% 0.00% init
> 249 root 85 0 7592K 0K select 0:00 0.00% 0.00% <sshd>
> 421 agcarver 85 0 3448K 0K wait 0:00 0.00% 0.00% <bash>
> 402 root 85 0 3444K 0K ttyraw 0:00 0.00% 0.00% <getty>
> 374 root 83 0 3584K 0K kqueue 0:00 0.00% 0.00% <inetd>
More information about the questions
mailing list