Unfortunately no one can be told what fun_plug is - you have to see it for yourself.
You are not logged in.
Hi Forum,
I'm starting to get some temperature reading problem with my DNS323. I installed FW 1.05 and fonz 0.5 on USB. Everthing works fine. I don't have too much fancy stuff on it except: smb, nfs (not kernel version), sshd, fan control. As you can see, just basic stuff.
Recently, I found that my temperature reading is 0C (32F), and fan running at a fixed RPM, with status: 194 (normally, it's 0 or 2). Even HD is very hot. So I try to reboot it, it refuse to reboot. I shut it down, wait for a while, and restart it, it went ok. After few days, it get temperature reading problem again. I had to power it off, and power it on to fix it. And when I wrote a small cron job to monitor the temperature reading. This morning it's happened again.
Anybody used to have this problem? Hardware failure? or just setting? What's your suggestion?
Thanks
Noodle
Offline
A CH3SNAS user reports a similar problem here: http://www.aroundmyroom.com/2008/07/04/ … re-104rc4/
Do you have TWSI errors in dmesg as well?
I really don't know the cause of the problem, but I think the NAS doesn't like the fancontrol script. Try to disable the fun_plug fancontrol and check if this solves the problem.
Last edited by EnricoM (2008-07-10 21:52:50)
Offline
Thanks for reply, it looks like a kernel bug. Probably I need to disable fan control. Let's see.
Noodle
Offline
I have noticed the same problem, a simple restart fixes it but also sometimes it just goes away.
I am going to modify the fancontrol script to run the fan at 3500 rpm when the temperature reading is zero.
Offline
skydreamer wrote:
I am going to modify the fancontrol script to run the fan at 3500 rpm when the temperature reading is zero.
I *think* that it's not possible to set the fan speed when the temperature is zero. Both temperature and fanspeed commands seem to run over the same TWSI (Two Wire Serial Interface) and data registers attached to it. In my opinion the register logic needs to be reset. I have no idea if this is possible.
Is it possible that the fancontrol script tries to write a value to the TWSI to adjust the fanspeed directly after it reads a temperature? Or is there a delay (sleep) in the script between any temperature reads and fan control writes?
Last edited by EnricoM (2008-07-11 17:06:27)
Offline
I can confirm that this is the case- when the temperature reading is zero the fan speed cannot be controlled.
I guess Enrico that you might right, it could be caused by reading the temperature/writing fan speed sequence however I wonder if actually sending 0 when the temperature drops below treshold is not the trigger?
Offline
I am not familiar with the fan control script, so I can not confirm that. Maybe you can force to send 0's and check what happens.
Offline
This happened to me once, I modified my fancontrol after that so that the fan never turns off completely - 2000 RPM is just about silent anyways for me. Also I added some code to bump it up if the temperature is 0, but I don't think that has happened since.
I believe it happened due to the memory being full. I had fancontrol writing a log file to /var/log/ and this happened after the dns had been running for a long time. Since the log was written once a minute, the size could reach quite large. I changed the script as well so that it only writes to the log when the temperature changes (not every minute) and shortened what is written to the log. I haven't had the trouble since, but it may just be a coincidence.
Offline
I don't think log is an issue. Actually, I disabled fan control log, no log at all, but it still happens. 2000RPM is enough if your HD spin down, however when I was using HD, HD went really hot, because FAN speed cannot change anymore. Anyway, I'm using a crontab job to monitor temperature reading for every 10 minutes. I'm thinking to change the script if temperature becomes 0, after 30 minutes, I will shutdown NAS. Hope this can be fixed.
BTW, I never this problem when I was using fw 1.04 and fonz 0.3. When you get this problem, which fw and fun_plug are you using? Could this be a bug in new firemware (1.05) ?
Offline
I have modified the fanman.sh to run the fan at 2,000 rpm on both enclosures instead of stopping it and will monitor the situation for a few days.
Old:
# Temp below T1 fan stops
CALCTEMP=`expr 0`
New:
# Temp below T1 fan stops
CALCTEMP=`expr 2000`
Offline
If there really is a bug that is causing the fan control script to fail, it might be prudent to move / remove the script on the wiki until the issue can be tracked down.
Offline
I don't think the bug is within fan control. It looks like bug in kernel that IO failure, but don't know what cause it.
Offline
FYI: I am running the modified fanctrl.sh script that does not turn off the fan but keeps it running at 2,000 r.p.m. and after one week the temperature reading is still OK on both DNS-323.
Offline
If you leave the fan running all the time at minimum 200 rpm, why install the fanspeed script? The standard firmware will keep it running also at minimum rpm when temperature is low.
Offline
knireis wrote:
If you leave the fan running all the time at minimum 200 rpm, why install the fanspeed script? The standard firmware will keep it running also at minimum rpm when temperature is low.
Oh well, then I can disable the script. I have modded the DNS-323 so extensively that the fan is barely noticeable even when running at a high speed.
Offline
Actually the fanman.sh has a softer temperature/speed curve so it still does its job even when it keeps the fan running at 2,000 rpm.
The embedded D-link control routine spins the fan to 4,000 rpm @42C, while the fanman is only idling at half the speed.
Offline
Any update on this error? Its happened pretty frequently now, almost once per day.
Offline
Lately I'm having the same issue. Temperature is 32F/0C. Fanspeed is around 2533rpm. The command fanspeed w 4000 does not change the fanspeed to 4000. I have ffp installed but without a fancontrol script. Lies the problem with the FFP or is it my CH3SNAS?
Offline
DeLaCroix wrote:
CH3SNAS: This error seems to be fixed in 1.04RC6, so be patient for the final release Thats an error in the Firmware which occurs from time to time.
I'm running 1.04RC6 already, so the problem still exists.
Offline
I've had my DNS-323 1.05 for about a week now running two 1TB drives in RAID 1. Temperature sensing seemed fine upon install and for a few days afterward.
I installed ffp and was using dns323-temp, and dns323-rtc to check both from the command line.
Now the dns323-temp is returning "0 C 32 F" and dns323-rtc is returning bogus values of "Mon Jan 1 01:41:41 2121".
The malfunction persists after reboot. Seems like it's keeping normal time OK and my fan seems to be running normally.
I recently made 3 changes which might have contributed:
1. setup NTP
2. setup TZ variable on boot
3. ran mediatomb (which I later killed because I don't need a uPNP server right now)
I have a suspicion that mediatomb might have been running and filling up the flash memory, now causing this problem. Is there anyway I can check my "available" flash memory?
Offline
i never used mediatomb and have the problem as well, so i guess it's not mediatomb
Offline
knireis wrote:
i never used mediatomb and have the problem as well, so i guess it's not mediatomb
My fan reports the exact same speed, 2533 rpm. What does dns323-rtc report for you?
I was suspecting mediatomb because I thought it might have filled my flash with logs. Did you setup/run any other application recently which might have done something similar? Have you setup NTP or TZ?
Offline
2546, it does not change much either, the temperature can increase quite a bit on the outside by heavy use. I removed the front panel for better airflow and that helps keeping it cool.
i have a ch3snas and have setup ntp, and use the fancontrol from here: http://nas-tweaks.net/CH3SNAS
But as i remember well, even without this fancontrol this problem occurs.
Offline
I think I may have a workaround for this issue.
Instead of polling the DNS-323 sensor temperature, which really does not bother me that much I have changed the fanman.sh script to get the internal temperature of the disks that is rather more important.
Ta=`/mnt/HD_a2/ipkg/opt/sbin/smartctl -d marvell -A /dev/sda | awk '/194 Temperature_Celsius/{print $10}' | sed 's/^[0]\+//'`
Tb=`/mnt/HD_a2/ipkg/opt/sbin/smartctl -d marvell -A /dev/sdb | awk '/194 Temperature_Celsius/{print $10}' | sed 's/^[0]\+//'`
if [ $Ta -gt $Tb ]; then
# Assign the higher temperature
T=$Ta
else
T=$Tb
fi
I am going to test this over a period of time to see if the problem with failing fan controls gets resolved.
You need to install ipkg and smartctl for this to work.
Offline