NAS 326 problem

MichalP

I have four ZYXEL 326 NAS. In different companies.One by one, they all stopped working properly. Within half a year.There is no access to them. The NAS is not responding to the network. There is no PING. Not over SSH either. Only reboot works for 3-4 days.I don't have TwonkyServer running which was using 100% CPU.

I have the impression that Zyxel set it up so that it is supposed to break at this time.Can anyone help? Not good to throw away the equipment.

Mijzelf

If you reboot them, how do you do that? If it reacts normally on the power button, the OS is still running. If you have to keep it pressed untill it's off, then the OS was crashed/vanished.

Is there something which was added to all 4 networks, last year? Are the NASses accesible from internet?

MichalP

Disconnecting the cable is only effective. Turning off with the power button does not work. Holding 10 seconds or more too.I had twonky and backups at home.The other three are just backups. So everything unnecessary was turned off. Various locations.Each NAS was set with a fixed IP. I have not accessed any of them from the internet.

Mijzelf

That is strange. Virtually every device which has a simple pushbutton power switch, has a dedicated microcontroller (MCU) to switch on/off the power. To switch the power on, the CPU and OS obviously cannot be used, as they are off. To power off, the 'push-event' is passed to the OS, which can shutdown gracefully. But if you keep the power button pressed, the MCU detects that, and simply cuts the power.

So if your box doesn't react on keeping the power button pressed, the MCU is crashed, or the button is defective. (When the button signal is not constant, the MCU might never kick in). Having said that, I am not in the posession of a 326, and so I can't be sure that functionality is programmed in the MCU. It is in the 325 and the 540. (And in every laptop I ever tested). Maybe worth to test on a working 326.

AFAIK on a 326 the MCU is also responsible for the fan. You can request the fan speed from the OS, but not change it.

I cannot really believe it's a ZyXEL set up. Where is the win? If you buy a NAS, and it doesn't live as long as you expected, then you won't buy that brand again, would you?

The only parts in a NAS which are susceptible to wear (apart from disks and fan) are the capacitors on the main board and in the power supply. But I can't think of a scenario where a box with defect capacitors can run for 2 or 3 days, and then fail. Unless the box crashes at disk spinup. The voltage might drop that much that the box freezes. Although I still would expect the MCU to carry on.

Could you as a test disable the disk spin down?

MichalP

It's all weird.Each of the four devices had the same symptoms.The one I have at home works for about 3-4 days, the ones I had at work up to two weeks. But the symptom is the same. Restart only by power cut.

I've already replaced it at work with another company and it works beautifully.But it's a pity at home because it's a bit of money.

You gave me some clues.First, replace the power supply.

Then maybe disable disk sleep after idle. And then the work of the equipment without disks.

The strange thing is that I have the device set to restart at four in the morning. When it freezes, it won't even do that.Sorry for the language but I use google translate

HappyCamper

I have the exact same problem. The push button is unresponsive. It wouldn't matter if I held it in for a month. The only way that I can restart these is to pull the power cord momentarily. I too have four (4) of these servers (326). I might have #1 go offline one week and then it could go for 3 weeks w/out any further incidents. Then, when the next one fails, it might be server #2, #3 or #4. There's no rhyme or reason. The only thing that's consistent is that I am forced to pull the power cord since the push button does not respond. Everything else on the networks works as it should. I sure wish that someone had some answers. My guess is that the servers are simply not well thought out from a software standpoint.

MichalP

Hi.I managed to partially solve it, although in most cases I had already replaced the hardware with synology.

When the equipment started to jam, I set it to automatic reset every day. As soon as I turned off the automatic daily reset, the equipment has been holding for 2 months without suspension.

franzk

Hi @Mijzelf,

I now have this behavior too (on 3 devices at the same time!). I recently installed the latest firmware on all devices. An automatic restart was defined on all devices on the first Friday of the month. Today is the first Saturday in the first month after the firmware update. 🙈

All devices are difficult or even impossible to reach via the LAN. Continuous PINGs keep breaking off and logging in via the website is rarely, if ever, possible. A restart is only possible by interrupting the power supply. No indication of defective HDD - neither on the device, nor in the admin page (if you ever get to it). Big catastrophe.

The push button to switch off via MCU has never worked on my devices (3 x 326), as stated in the operating instructions. At least 2 were bought "new" in stores. I always used the admin website to shut down or restart. Because of the price of the device, I simply “accepted” this as given.

Synology is significantly better - but also more expensive!
If I don't find a solution in the next few days (firmware downgrade), I will unfortunately have to replace all devices.

MichalP

MichalP

https://community.zyxel.com/en/discussion/comment/60546#Comment_60546

I have already replaced most of my equipment with Synology.

But I still left Zyxel at home. I noticed that the problem is the automatic restart of the device. After turning off automatic restart, I have had peace of mind for half a year and the device has not broken down.

And this is a vicious circle. The more the device broke down, the more often I set automatic restart. And so on.

Mijzelf

@MichalP That is interesting. As the problem seems connected with network, I suppose that means that the NIC doesn't reinitialize well when rebooted without powerdown in between. But that doesn't explain why the MCU shutdown also doesn't work.

If you are still interested, there might be a work-around. The hardware supports 'wake on RTC', yet I don't know if the firmware does. Meanwhile I have a 326 running Debian, which switches itself on twice a week, performs a backup, and switches itself off again. The initiation of this is basically:

echo `date '+%s' -d '01:00 next tuesday'` > /sys/class/rtc/rtc0/wakealarm
poweroff

This injects the unix time of '01:00 next tuesday' into the RTC, which will power up the NAS at that time. This works fine on Debian, the box has never missed a wakeup moment, nor did it ever fail to perform it's task. Possible problems on the stock firmware:

1 Don't know if the RTC driver supports it this way.

2 Don't know if 'date' (on the 326 a busybox applet) supports this. If not, you could install Entware-ng, which can provide the full GNU version of date.

Of course you can also install full Debian (with a recent kernel) on that box. AFAIK it's rock stable, the only drawback I know is that samba access is slower. (Maybe due to an over-optimized samba on the firmware). (And of course you loose the firmware webinterface. Administration has to be done on the commandline, just like every headless Linux server). Instructions an forum here.