Hi, We have a HP R/T 3000 G4 UPS; according to the data returned by the usbhid-ups driver, it seems to be a "badge engineered" Eaton 5PX UPS LI R 3000. We have a stubborn "Replace Battery" alarm that stays despite: - physical battery replacement - successful quick battery test, started by upscmd - successful deep battery test, started by upscmd - successful battery test, started from panel - "reset fault state" on panel - clear fault log on panel The panel doesn't show any battery-related alarm. I initially "guessed" this was some kind of stale flag in some internal NUT database or cache, but I couldn't find on the hard disk where that is/was, and restarting all NUT daemons & drivers didn't help either. Anybody has a clue how I should proceed or debug the issue? Thanks in advance, Lionel Mamane
On Mon, 21 Sep 2020, Lionel Élie Mamane wrote:> We have a HP R/T 3000 G4 UPS; according to the data returned by the > usbhid-ups driver, it seems to be a "badge engineered" Eaton 5PX UPS > LI R 3000. > > We have a stubborn "Replace Battery" alarm that stays despite: > ... > The panel doesn't show any battery-related alarm. > > I initially "guessed" this was some kind of stale flag in some > internal NUT database or cache, but I couldn't find on the hard disk > where that is/was, and restarting all NUT daemons & drivers didn't > help either.The upsd daemon keeps an image of the current state of the UPS unit in memory. There is no external database. The monitor daemon upsmon polls this image.> Anybody has a clue how I should proceed or debug the issue?At first sight this looks like a "feature" of the particular UPS unit you have, but perhaps other have deeper insight. Roger
On 9/21/20 9:47 PM, Lionel Élie Mamane wrote:> Hi, > > We have a HP R/T 3000 G4 UPS; according to the data returned by the > usbhid-ups driver, it seems to be a "badge engineered" Eaton 5PX UPS > LI R 3000. > > We have a stubborn "Replace Battery" alarm that stays despite: > > - physical battery replacement > - successful quick battery test, started by upscmd > - successful deep battery test, started by upscmd > - successful battery test, started from panel > - "reset fault state" on panel > - clear fault log on panel > > The panel doesn't show any battery-related alarm. > > I initially "guessed" this was some kind of stale flag in some > internal NUT database or cache, but I couldn't find on the hard disk > where that is/was, and restarting all NUT daemons & drivers didn't > help either. > > Anybody has a clue how I should proceed or debug the issue?try to run the driver in debug mode so that we can see what data is actually sent by the UPS #usbhid-ups -DDDD -u root -a $YourUpsName where $YourUpsName is the one you used in /etc/ups/ups.conf After a few seconds the output will start to repeat so ctrl-c and examine whatever was collected. and maybe zip it and send it to the list.
On Thu, Sep 24, 2020 at 06:51:23PM +0300, Manuel Wolfshant wrote:> On 9/21/20 9:47 PM, Lionel Élie Mamane wrote:>> We have a HP R/T 3000 G4 UPS; according to the data returned by the >> usbhid-ups driver, it seems to be a "badge engineered" Eaton 5PX >> UPS LI R 3000.>> We have a stubborn "Replace Battery" alarm that stays (...):>> The panel doesn't show any battery-related alarm.>> Anybody has a clue how I should proceed or debug the issue?> try to run the driver in debug mode so that we can see what data is actually > sent by the UPS> #usbhid-ups -DDDD -u root -a $YourUpsName> where $YourUpsName is the one you used in /etc/ups/ups.conf> After a few seconds the output will start to repeat so ctrl-c and examine > whatever was collected. and maybe zip it and send it to the list.I attach the output. I looked for "replace" and "battery" in there, and nothing obviously understandable to me in there; that is too raw data, not sure what it means. If anybody can make sense of it, thanks in advance! Lionel -------------- next part -------------- A non-text attachment was scrubbed... Name: usbhid-ups_output.bz2 Type: application/octet-stream Size: 17871 bytes Desc: not available URL: <http://alioth-lists.debian.net/pipermail/nut-upsuser/attachments/20200925/e15474c1/attachment-0001.obj>
On Tue, Sep 22, 2020 at 08:28:46AM +0200, Roger Price wrote:> The upsd daemon keeps an image of the current state of the UPS unit > in memory. There is no external database. The monitor daemon upsmon > polls this image.So I understand that restarting upsd is guaranteed to reset that image.