DSM-G600, DNS-3xx and NSA-220 Hack Forum

fonz · 2008-10-28 16:15:22

skydreamer wrote:
Hi Fonz, What is the best way to upgrade to the new kernel? I am a bit concerned about the whole temperature issue.

You can't run the firmware with vanilla kernels, so you need to install something else, e.g. ffp-reloaded, ARMedslack or Debian.

jesbo · 2008-10-28 16:38:52

FYI... I installed the fanctl binary and now start it automatically at bootup. I am seeing occasional messages in dmesg:

TWSI: twsiDataTransmit ERROR - Read Data Int TimeOut.

However, these do not seem to be affecting ability to read temperature or set fan speed (at least not so far). Fanctl still seems to be working OK. I can independently use "temperature g 0" and "fanspeed g" to see that things seem normal. The DNS-323 has been booted for at least 3 days now with no issue.

Papaloewe · 2008-10-30 15:43:55

Maybe I have found the solution für the TWSI-Errors in this blog entry (german):
http://wolf-u.li/1802/firmware-104rc6-f … rschienen/

"...Ich habe die Vermutung das es an dem ntpd Dienst liegt. Ich habe ihn jetzt deaktiviert und nun läuft das NAS. Mir ist aufgefallen, dass der drift (/ffp/etc/ntp.drift) trotz der Korrektur der Ticks im /ffp/etc/fun_plug.local die 500.000 erreicht. Dieses könnte erklären warum der Fehler erst nach ein paar Tagen auftritt. Habe mir nun einen cronjob hinzugefügt der mir stündlich die Zeit mit sntp einstellt. Mal sehen was die nächste Woche ergeben wird. fanctl läuft soweit super..."

The user seb suggested that the ntpd deamon from the fun_plug is the causer for the TWSI-errors.

I tried it out and bingo. Since I put it off:
"chmod -x /ffp/start/ntpd.sh"

No more TWSI-errors appeared.

Last edited by Papaloewe (2008-10-30 15:51:56)

jesbo · 2008-10-30 15:53:57

Interesting. I do run the ntpd process.

knireis · 2008-10-30 17:19:45

jesbo wrote:
Interesting. I do run the ntpd process.

Me too......, well not anymore, switched it off after reading the news. See how it goes from now.

jesbo · 2008-10-30 17:32:27

knireis wrote:
jesbo wrote:
Interesting. I do run the ntpd process.
Me too......, well not anymore, switched it off after reading the news. See how it goes from now.

I have not turned it off. Since I changed from running the infinite loop "temperature g 0" / "fanspeed w n" shell script to the fanctl binary, I have seen only 1 of these in the log:

TWSI: twsiDataTransmit ERROR - Read Data Int TimeOut.

My DNS has been up and running for about a week now with both fanctl and ntpd running. I have not lost the ability to control the fan.... yet.

Not to go totally off topic... but has anyone noticed the "uptime" output is bogus? Whats that all about:

/ # uptime
11:36:29 up 14182 days, 15:36, load average: 0.00, 0.00, 0.00
/ #

Last edited by jesbo (2008-10-30 17:37:52)

rolfsky · 2008-10-30 17:38:34

knireis wrote:
jesbo wrote:
Interesting. I do run the ntpd process.
Me too......, well not anymore, switched it off after reading the news. See how it goes from now.

And I run it as well... maybe we have found a commonality.

However, the translated German is complaining about the ntp.drift file veering off towards 500 ppm (the upper bound of ntpd's non-step corrections.) I have now adjusted my DNS-323 so that the drift is consistently within bounds at < 300ppm. (9967 ticks, but as fonz says, ymmv)

(btw, 1 tick is roughly 100ppm, fonz's default value of 9965 brought me to about 330ppm in the drift, 9966 ~ 230ppm and 9967~130ppm)

Everything had been going fine since my last post, until I adjusted the ticks again to 9967. Then within a day or so I was getting the temp reading error. Related? A full shutdown was required to remedy the problem.

Perhaps the issue isn't with ntp, but rather with adjtimex or related?

perssinaasappel · 2008-10-30 17:57:43

I was also thinking to the adjtimex statement.
My ticks are now 10010 which results in '0' for the ntp.drift file.
At this time no TWSI errors.

I will change the adjtimex value to something else to see if the TWSI errors come back.

Little off topic, after a while I don't see new ntp.drift files generated while the deamon is still running.
Also /var/log/messages doesn't show adjustments.

Weird.

Last edited by perssinaasappel (2008-10-30 17:59:47)

jesbo · 2008-10-30 18:10:10

I am using the base clock value of 9965 and am seeing ntp.drift stable at 335. Have not messed with it further since the system clock seems accurate now... at least enough for my needs. My ntp.drift file seems to be updated fairly regularly, but the value in it doesn't change much. Last update it went from 335.05 to 335.029.

Last edited by jesbo (2008-10-30 18:12:37)

rolfsky · 2008-10-30 18:54:15

perssinaasappel wrote:
I was also thinking to the adjtimex statement.
My ticks are now 10010 which results in '0' for the ntp.drift file.
At this time no TWSI errors.

I will change the adjtimex value to something else to see if the TWSI errors come back.

Little off topic, after a while I don't see new ntp.drift files generated while the deamon is still running.
Also /var/log/messages doesn't show adjustments.

Weird.

Hi perssinaasappel, your comments about the ntp.drift not being generated while the daemon is running and also a value of 0 make me a little curious. Also, how did you come about the value of 10010 for ticks?

Can you run ntpq -c rv and tell us what the "state=" value is? If you have a state != 4, then ntpd is probably not running correctly.

Thanks!

perssinaasappel · 2008-10-30 20:12:55

Hi rolfsky, I did some experiments with the adjtimex value.
Changed it little by little and looked in the ntp.drift file to see what happens.
(After change let it settle down a few days to prevent over adjusting.)
This way I ended up with 10010 ticks and ntp.drift with '0'.
Like fonz said, ticks = 9965, your mileage may vary.
FYI, adjtimex accepts values from 9000 - 11000.

I'll give you the output of 'ntpq -c rv' when it seems stuck.
I just rebooted the darn thing because I changed the ticks value for this TWSI experiment.
<edit> Hmmm, after reboot it gives me directly state=4, is this normal? </edit>

Last edited by perssinaasappel (2008-10-30 20:18:07)

knireis · 2008-10-30 23:09:42

I switched off the ntpd deamon earlier this afternoon, but unfortunately the temperature is back to 0/32. So i guess it must be something else.

rolfsky · 2008-10-30 23:14:40

knireis wrote:
I switched off the ntpd deamon earlier this afternoon, but unfortunately the temperature is back to 0/32. So i guess it must be something else.

As a stop-gap, could we modify the existing "over-temperature" email warning to also alert for "under-temperature"?

Because if my temp is unregulated, the sooner I know the better.

perssinaasappel · 2008-10-30 23:16:04

@knireis: Did you remove the 'adjtimex' statement from the fun_plug.local file?
And maybe 'timezone' and 'timeserv' statements?

Last edited by perssinaasappel (2008-10-30 23:17:27)

rcblackwell · 2008-10-30 23:39:09

Thw DNS323 here beside me has been running without a fanctl script for twelve plus days. When the script was first disabled DMESG returned no TWSI errors. I checked tonight and was surprised to see DMESG is full of TWSI messages. I'm running the ntpd.sh script and set ticks during the boot process. I'm going to disable the script and tciks to see what happens.

knireis · 2008-10-31 08:26:51

perssinaasappel wrote:
@knireis: Did you remove the 'adjtimex' statement from the fun_plug.local file?
And maybe 'timezone' and 'timeserv' statements?

No, I did not.
1 minute later: now i have removed the adjtimex from the funplug.local, see what this brings. I'll wait with the timezone/timeserv.

knireis · 2008-10-31 16:22:06

No more 0/32 temperature readings, sofar. I've put this line in the editcron.sh file to adjust the time to a timeserver once a day:
/bin/echo "5 1 * * * /usr/sbin/sntp -r de.pool.ntp.org >> /ffp/log/sntp.log" >> $CRONTXT

perssinaasappel · 2008-10-31 17:31:16

Looks promising here also.
If it really is the adjtimex statement can this be explained by the guru's?

I'll let it run for a few more days to be sure.

rcblackwell · 2008-10-31 22:59:31

rcblackwell wrote:
Thw DNS323 here beside me has been running without a fanctl script for twelve plus days. When the script was first disabled DMESG returned no TWSI errors. I checked tonight and was surprised to see DMESG is full of TWSI messages. I'm running the ntpd.sh script and set ticks during the boot process. I'm going to disable the script and tciks to see what happens.

I know have a DST problem. The DNS is running one hour ahead of the current time!

The following lines have been added to the fun_plug.local file yet the time is still of by an hour.

Code:

# Canada Eastern Standard Time
echo 'EST5EDT,M3.2.0,M11.1.0' >/etc/TZ

# Calibrate time
/usr/sbin/sntp -r time.nrc.ca

I've also added the following crontab

Code:

"30 4 * * * /usr/sbin/sntp -r time.nrc.ca >> /ffp/log/sntp.log"

Any help by way of a temporary solution while the ntpd.sh script and adjtimex function is being investigated as a casue of TWSI errors would be appreciated.

DaveN · 2008-11-01 05:38:25

I also have the temperature reading problem.

I'm running DNS-323; FW 1.05; ffp 0.5 (telnet, ntpd, plus fanctl from wolf.li and nightly cron jobs to backup from vol A to B) mounted from a USB stick.

I'm interested in the theory that ntpd and/or adjtimex has something to do with the problem. Right now my box has a drift of 500, the temperature reads 0 and dmesg is full of TWSI errors. None of these is normal.

I think that the problem may exist even without the fanctrl script. I have no proof as I didn't check the temperature reading or fan speed (or dmesg) when I noticed an earlier problem with the drift, however the drift symptoms were identical.

In late September, I had the same setup above, but with no fanctl. I had just finished tweaking adjtimex (to 9968) so that the drift was around 17~18 ppm. I turned on statistics logging (to a file on the usb drive using ntp.conf: statistics loopstats peerstats) as I was interested in determining how the drift would change over time. The drift climbed very slowly, but fairly steadily between Sept 27 (17~18 ppm) and Oct 5 (36~37 ppm). From Oct 5 to Oct 14 the drift stayed steady in this range, then on Oct 16 the drift went wild. See the log file below:

loopstats (day, second, offset, drift compensation, estimated error, stability, polling interval)
54754 172.321 -0.003604000 37.417 0.025719229 0.043243 10
54754 2236.277 -0.013677000 37.417 0.024320283 0.040450 10
54754 3267.243 0.022054000 37.501 0.026021599 0.048254 10
54754 4299.223 0.006182000 37.525 0.024979440 0.045913 10
54754 6361.178 -0.004484000 37.525 0.023668441 0.042947 10
54754 7391.139 -0.023179000 37.436 0.023105356 0.051014 10
54754 8424.117 0.021265000 37.518 0.026721267 0.055802 10
54754 9454.085 0.007207000 37.546 0.025484819 0.053106 10
54754 10486.063 -0.005822000 37.523 0.024279820 0.050302 10
54754 12549.010 -0.023691000 37.523 0.023574028 0.047053 10
54754 13579.973 0.021493000 37.606 0.027229886 0.052808 10
54754 15642.940 -0.009237000 37.606 0.027691526 0.049398 10
54754 16675.921 -0.018764000 37.534 0.026121159 0.052790 10
54754 18738.861 0.016202000 37.534 0.027383479 0.049381 10
54754 20801.814 -0.013495000 37.534 0.027683321 0.046191 10
54754 22865.774 -0.022802000 37.534 0.026103591 0.043208 10
54754 23896.740 -0.032176000 37.410 0.024641569 0.059516 10
54754 25959.673 0.009006000 37.410 0.027263598 0.055672 10
54754 26989.638 -0.011622000 37.365 0.026525078 0.054410 10
54754 28020.605 -0.018872000 37.293 0.024943992 0.056983 10
54754 30083.542 0.016064000 37.293 0.026400703 0.053303 10
54754 32146.497 -0.015450000 37.293 0.027092721 0.049860 10
54754 34211.461 0.043881000 37.293 0.032898056 0.046640 10
54754 35378.089 0.000000000 37.293 0.000007629 0.043627 4
54754 35416.422 0.038460000 38.687 0.013597586 0.494458 6
54754 35421.443 0.113878000 39.230 0.029542612 0.500783 6
54754 36373.542 0.000000000 39.230 0.000007629 0.468440 4
54754 36394.706 0.087322000 40.979 0.030872966 0.757825 6
54754 36410.819 0.053408000 41.793 0.031269232 0.765198 6
54754 36460.162 0.091835000 46.172 0.032250945 1.705674 6
54754 37369.998 0.000000000 46.172 0.000007629 1.595512 4
54754 37385.131 -0.007848000 46.053 0.002774621 1.493065 6
54754 37449.614 0.062225000 49.851 0.024909944 1.937422 6
54754 37593.558 0.101509000 63.791 0.027126602 5.251223 6
54754 38372.479 0.000000000 63.791 0.000007629 4.912069 4
54754 38382.551 0.001506000 63.805 0.000532363 4.594823 6
54754 38407.730 -0.011040000 63.542 0.004463423 4.299070 6
54754 38456.054 0.042397000 65.523 0.019348635 4.081960 6
54754 38555.755 0.035867000 68.909 0.018245656 4.001626 6
54754 38813.530 0.122104000 98.953 0.034941291 11.262213 7
54754 39655.983 0.000000000 98.953 0.000007629 10.534835 4
54754 39678.149 -0.007198000 98.795 0.002544805 9.854594 6
54754 39692.208 0.076915000 99.822 0.029833521 9.225276 6
54754 39710.363 0.050324000 100.686 0.029447727 8.634859 6
54754 39736.623 0.032208000 101.484 0.028280663 8.082104 6
54754 39969.171 0.023954000 106.807 0.026614604 7.790814 7
54754 40678.599 0.000000000 106.807 0.000007629 7.287639 4
54754 40702.789 0.061901000 108.224 0.021885355 6.835341 6
54754 40747.146 0.054767000 110.574 0.020626668 6.447648 6
54754 41672.018 0.000000000 110.574 0.000007629 6.031222 4
54754 41699.222 -0.027116000 109.876 0.009587100 5.647090 6
54754 41741.497 0.032977000 111.197 0.023061456 5.302972 6
54754 41742.513 0.031402000 111.227 0.021579207 4.960487 6
54754 41856.302 0.096636000 121.733 0.030649476 5.943738 6
54754 41920.765 0.105892000 128.196 0.028856138 6.011118 7
54754 43151.973 0.000000000 128.196 0.000007629 5.622886 4
54754 43168.104 0.075003000 129.412 0.026517654 5.277268 6
54754 43203.312 0.048293000 131.024 0.026541850 4.969221 6
54754 43628.308 0.124508000 181.489 0.036640183 18.437463 6
54754 44337.824 0.000000000 181.489 0.000007629 17.246667 4
54754 44349.928 0.010485000 181.609 0.003707039 16.132836 6
54754 44384.123 0.055078000 183.447 0.016142877 15.104877 6
54754 44500.993 0.045359000 188.465 0.015486343 14.240261 6
54754 45344.352 0.000000000 188.465 0.000007629 13.320545 4
54754 45381.626 0.069524000 190.918 0.024580278 12.490380 6
54754 45443.028 0.035121000 192.995 0.026011686 11.706726 6
54754 45640.454 0.123567000 216.210 0.039621583 13.685156 6
54754 46227.646 0.117386000 281.923 0.037126972 26.526507 6
54754 46616.807 0.000000000 281.923 0.000007629 24.813276 4
54754 46626.856 0.012493000 282.042 0.004417100 23.210732 6
54754 46633.939 0.004089000 282.070 0.005089347 21.711654 6
54754 46683.296 0.029765000 283.489 0.010250435 20.315591 6
54754 46732.644 0.047625000 285.714 0.011480876 19.019778 6
54754 47256.362 0.067085000 319.238 0.012754299 21.377928 6
54754 47779.531 0.000000000 319.238 0.000007629 19.997220 4
54754 47795.665 -0.001621000 319.214 0.000573134 18.705689 6
54754 47808.714 -0.023847000 318.918 0.007876285 17.497882 6
54754 47827.879 -0.034694000 318.289 0.008305932 16.369279 6
54754 47862.125 0.003751000 318.415 0.015656076 15.312122 6
54754 48519.764 0.085023000 371.687 0.032251050 23.662126 6
54754 49103.378 0.000000000 371.687 0.000007629 22.133892 4
54754 49120.580 0.004546000 371.761 0.001607154 20.704377 6
54754 49128.570 -0.000040000 371.760 0.002210982 19.367171 6
54754 49171.902 0.015844000 372.410 0.005984465 18.117786 6
54754 49487.136 0.067906000 392.874 0.019239022 18.427437 6
54754 50005.893 0.079125000 431.962 0.018428452 22.093168 6
54754 50135.855 0.083226000 442.280 0.017299093 20.985772 6
54754 50592.390 0.000000000 442.280 0.000007629 19.630392 4
54754 50621.618 0.053417000 443.758 0.018885636 18.369977 6
54754 50674.004 0.067979000 447.129 0.018400849 17.224826 6
54754 50737.475 0.063421000 451.000 0.017287687 16.170368 6
54754 50867.433 0.054434000 457.748 0.016480319 15.313022 6
54754 51383.201 0.059669000 487.111 0.015526625 17.690388 7
54754 52225.622 0.000000000 487.111 0.000007629 16.547842 4
54754 52232.695 0.014128000 487.206 0.004995129 15.479125 6
54754 52234.720 0.012653000 487.230 0.004701552 14.479398 6
54754 52290.123 0.012288000 487.886 0.004399784 13.546224 6
54754 52598.344 0.001382000 488.292 0.005639674 12.672145 6
54754 52983.124 0.024536000 497.301 0.009738525 12.274161 7
54754 53369.133 0.000000000 497.301 0.000007629 11.481427 4
54754 53374.190 -0.001081000 497.296 0.000382235 10.739891 6
54754 53441.635 0.015225000 498.268 0.005776135 10.052134 6
54754 53554.498 0.023132000 500.000 0.006083419 9.444125 6
54754 53618.969 0.024663000 500.000 0.005716222 8.870437 6
54754 54205.235 0.025783000 500.000 0.005361658 9.319285 6
54754 54269.708 0.033178000 500.000 0.005656001 9.694996 6
etc... (from here on until I rebooted the drift was 500).

If I compare my logstats file from yesterday it is very similar - and today I have the drift of 500, the temperature reads 0 and dmesg is full of TWSI errors.

Any comments on the above?

Now back to the topic of the original post:
In trying to create my own fancontrol.sh script earlier this week I experienced the temperature problem immediately after running a script containing a coding error -effictively my while loop did not have a sleep so the script continually read the temperature over and over again. I'll try to reproduce this error to see if I can trigger the temperature problem reproducibly. If I can then I'll start investigating with and without ntpd/adjtimex etc.

Cheers,
Dave.

fonz · 2008-11-01 14:22:17

DaveN wrote:
In trying to create my own fancontrol.sh script earlier this week I experienced the temperature problem immediately after running a script containing a coding error -effictively my while loop did not have a sleep so the script continually read the temperature over and over again.

Clock, fan and temp sensors share the TWSI. Maybe it's simultaneous accesses to the TWSI that creates the problems.

DeLaCroix · 2008-11-01 14:49:55

fonz wrote:
Maybe it's simultaneous accesses to the TWSI that creates the problems.

Makes sense as (what i read) all people with deactivated time-services and activated fanctl do not have this problems. Perhaps the time-services should be called from fanctl or even included into fanctl to have a ordered access to TWSI?

knireis · 2008-11-07 12:02:03

To get rid of the temp raeding problems i disabled the ntpd, errors still returned. Then i also disabled the adjtimex in funplug.local, with succes, no more temp reading problems. I read somewhere that changing the adjtimex value from 9965 to 9960 also helps. I tried that and guess what, no more errors either. Any clues why 9965 gives errors and 9960 does not?

jesbo · 2008-11-07 16:25:17

I am running my fanctl script with checking done every 300 seconds (5 minutes seemed adequate to me). My adjtimex is set at 9968 and I'm seeing a drift in ntp.drift that hovers around 32. I am no longer seeing any TWSI errors. Things seem to be running pretty smoothly.

perssinaasappel · 2008-11-07 20:42:21

Just for the record,

After removing the adjtimex statement no more TWSI errors on my side. (Default ch3snas fan control.)
(Also removed the ntpd daemon and now using extra cron job.)

I added a line in editcron.sh to adjust time once every day @ 6:55 am. Works great for me.

Code:

/bin/echo "55 6 * * * /usr/sbin/sntp -r pool.ntp.org >> /ffp/log/sntp.log" >> $CRONTXT

Last edited by perssinaasappel (2008-11-07 20:52:57)

DSM-G600, DNS-3xx and NSA-220 Hack Forum

Announcement

#51 2008-10-28 16:15:22

Re: Temperature reading problem!

skydreamer wrote:

#52 2008-10-28 16:38:52

Re: Temperature reading problem!

#53 2008-10-30 15:43:55

Re: Temperature reading problem!

#54 2008-10-30 15:53:57

Re: Temperature reading problem!

#55 2008-10-30 17:19:45

Re: Temperature reading problem!

jesbo wrote:

#56 2008-10-30 17:32:27

Re: Temperature reading problem!

knireis wrote:

jesbo wrote:

#57 2008-10-30 17:38:34

Re: Temperature reading problem!

knireis wrote:

jesbo wrote:

#58 2008-10-30 17:57:43

Re: Temperature reading problem!

#59 2008-10-30 18:10:10

Re: Temperature reading problem!

#60 2008-10-30 18:54:15

Re: Temperature reading problem!

perssinaasappel wrote:

#61 2008-10-30 20:12:55

Re: Temperature reading problem!

#62 2008-10-30 23:09:42

Re: Temperature reading problem!

#63 2008-10-30 23:14:40

Re: Temperature reading problem!

knireis wrote:

#64 2008-10-30 23:16:04

Re: Temperature reading problem!

#65 2008-10-30 23:39:09

Re: Temperature reading problem!

#66 2008-10-31 08:26:51

Re: Temperature reading problem!

perssinaasappel wrote:

#67 2008-10-31 16:22:06

Re: Temperature reading problem!

#68 2008-10-31 17:31:16

Re: Temperature reading problem!

#69 2008-10-31 22:59:31

Re: Temperature reading problem!

rcblackwell wrote:

Code:

Code:

#70 2008-11-01 05:38:25

Re: Temperature reading problem!

#71 2008-11-01 14:22:17

Re: Temperature reading problem!

DaveN wrote:

#72 2008-11-01 14:49:55

Re: Temperature reading problem!

fonz wrote:

#73 2008-11-07 12:02:03

Re: Temperature reading problem!

#74 2008-11-07 16:25:17

Re: Temperature reading problem!

#75 2008-11-07 20:42:21

Re: Temperature reading problem!

Code:

Board footer