UPS problem. Install regular NUT server?

[nickba]

Hello Fortran, yes I saw that. But in my case the error is always 71. I never got a 110.

Telos, my issue here is that I decided to move all my services to VMs on Synology, so I would not need to run anything else on an separate hardware. The only thing that is bugging me is this ups issue. I think there is a lot of people with the same issue, but they actually never noticed it, as the only thing that happens is that the DSM stops refreshing ups data, but the DSM never reports the ups was disconnected. In case of a power failure, DSM will just ignore it and the Synology will not shutdown gracefully and the whole purpose of the ups will be defeated.

 

[Robbie]

Depressing thread. This motivates me to move NUT to my RPi.

Same. Mine works just fine but after the early DSM 7 faffs with UPS connections and seeing the wider attempts to gimp open-source software to become vendor specific, it has made me consider moving my NUT to a different system.

I do have a nut-capable app running on my Mac mini server (mainly to shut the Mac down nicely) so I could use that. I have also considered using my router as the NUT host, which may be more logical yet does not get mentioned as much as a favoured option.

...but the DSM never reports the ups was disconnected.

This seems odd as I get CMS emailed-warnings as soon as there is the slightest hiccup. Different issue?

️

 

[Deleted member 5784]

I have also considered using my router as the NUT host, which may be more logical yet does not get mentioned as much as a favoured option.

Now that's a solution that I like. Always on, low power and inherently connected to everything that counts if/when there's a power issue.

This seems odd as I get CMS emailed-warnings as soon as there is the slightest hiccup. Different issue?

In this case, the Syno thinks it's blisfully connected to the UPS even when it isn't; hence, no notices generated. The errors about the comms failure are all occurring internally in the NUT driver, which doesn't bubble up events to the notification module.
Ofc This whole issue would be avoided if Syno didn't run such hugely outdated s/w. Their version of NUT is likely > a decade old by now
I'm running my 'real' NAS on Ubuntu 22.04LTS, and between removal of external USB functionality, misuse of SMART HDD reporting mechanisms to upsell oem drives and issues like this, i'm reminded almost daily what a good decision this was...

 

[Robbie]

Just to confirm, you are using CMS?

 

[Deleted member 5784]

Just to confirm, you are using CMS?

? CMS as in 'Content Management System'?
You've lost me there, i'm afraid.

EDIT: I guess you're talking about this; I have no direct experience, but it seems unlikely that it would notify of the UPS issue for the reasons I set out above, ie the Syno itself still thinks the UPS is connected & running. The issue the OP has is occuring within the NUT driver itself.

Centralised admin / error reporting tools like this tend to be about going wide (centralising lots of devices) rather than deep (monitoring services etc at a deeper level than a single machine would). I doubt that Syno's CMS is aware of issues this deep in the stack...?

 

[Robbie]

I get more reliable notifications with CMS.

 

[nickba]

Hello, I found an old Orange Pi(similar to Raspberry Pi) in my electronics stuff. Installed Armbian(Debian based) and nut. Adjusted it so nut server will only start when NAS is available. This is to avoid NUT shutting down the UPS if a power surge occurs during NAS boot up. Working perfectly!

Now the funny part: according to Synology, my UPS is defective LoL! Synology turned a great open source and reliable piece of soft like NUT in a completely unreliable closed source mess!

 

[Deleted member 5784]

Now the funny part: according to Synology, my UPS is defective LoL! Synology turned a great open source and reliable piece of soft like NUT in a completely unreliable closed source mess!

I think it's is the fact that they're using a very out of date version of NUT with long-patched bugs that's the problem here, rather than it being an open / closed source thing. But yes, they deserve to be called out for not updating their s/w stack. After all, the updated versions are aailable to them for free.

Glad you found a workaround to your issue.