DX4600，It used to display normally before, but after the update, nothing displays properly anymore

[kou147258]

[Tom-Bom-badil]

Which update? NAS Firmware or HA Integration?

• restart both HA and NAS,
• check on the NAS if the user you have specified is admin,
• reconfigure the NAS once (HA Settings menu).

/tom

[Tom-Bom-badil]

Weird. Any clue what the GPU has to do with this? The integration is not even polling GPU related entities so far.

The integration is using the same API the Web GUI and the Mobile App are using. If those are also not getting any data from the /stat/get_all/ endpoint, it's an indication to me that this is caused by UGOS. Also, @kou147258 mentioned this behaviour started after the UGOS June update.

So, I am not even sure that I can repair anything here. If we somehow are able to trace the problem down to the integration, I will be happy to repair the issue. But for the time being, I don't even know where to start ...

/tom

[de-goldi]

Thanks. Just checked and for me the problem started after upgrading to UGOS 1.17.0.0095 on July 7th. So could very well be a glitch in UGOS that was introduced with the new version. It's intermittently working after a NAS reboot, then stops working again after some random number of hours. Hard to debug. Let's see if the next update will fix it. Fingers crossed.

[Tom-Bom-badil]

@de-goldi Not sure if you are on the UGreen Discord, but it seems other people experiencing the same behaviour you reported: click. There have been a couple more 'memory leak' reports on that Discord over the last days - unrelated to this HA integration, more of a general UGOS issue.

/tom

[de-goldi]

@Tom-Bom-badil Thanks for letting me know! Haven't signed up yet for Discord but I guess know is the time to try it out.

[Tom-Bom-badil]

Quick update: Now there's a nice analysis of this in the linked Discord thread. Seems like a general UGOS issue; a corresponding report has been filed at UGREEN. Restarts seem to fix it for a while; also systemctl stop taskmgr_serv (although that is killing a lot more in the system, including API endpoints). Still not sure if this is related to @kou147258's original post, or something different.

/tom