On Aug 19, 2020, at 6:22 AM, Manuel Wolfshant wrote: > > On 8/19/20 12:55 PM, Marc Franquesa wrote: >> Thanks all for the quick responses on the question (always a good sign of >> this mailing list). >> >> (Tim Dawson, Manuel Wolfshant): >> Regarding root cause of my current problem pointing to a faulty battery, >> this is not the first time that someone points to it. However I made a >> simple test with the vendor tools (Windows) and didn't found any issue, as >> well as they discharged normally (although only tested without load). >> > That test is not very relevant, unfortunately. You can easily get false > results ( as in battery OK when it is not really OK ). The only relevant test > is under real load, which is why I suggested using a small motor or a light > bulb.
An UPS battery test (especially on a higher-end UPS like many of the MGE models) can do two things: provide a small load, and recalibrate the runtime estimate based on how the UPS responds to that load. The UPS may not result in a NUT "RB" (replace battery) flag, but it should still have a better estimate of runtime. Marc indicated he saw adequate runtime figures, but if those are based on the original battery, then they are bogus. >> >> (David Zomaya). >> Regarding Tripplite, I found many active threads on the mailing list so I >> consider it everytime as an alternative, however those devices are a bit >> pro/expensive too and they are harder to get/find support on my country >> compared with other vendors. However feel free to suggest a model like the >> Ellipse ECO 1200VA (in power and price). >> >> Also I found many reports of NUT working devices on the mailing list which >> are not reported on the HCL (https://networkupstools.org/stable-hcl.html) >> which makes me ask if this list is maintained/updated as well as the >> acknowledges section (https://networkupstools.org/acknowledgements.html) > > Unfortunately that list is rather unmaintained. The HCL on the website is built as part of the NUT release process, and due to a combination of blocking bugs (which I'm not going to rehash here; see GitHub), we haven't made a new release recently. Also note that if an UPS implements the PhoenixTec Qx protocol "well enough" for blazer_usb or nutdrv_qx, we don't often get reports of success. However, the HCL doesn't adequately capture a lot of the nuances of how *well* certain aspects of UPSes are supported by NUT. The Device Dump Library (DDL) has annotations on what is known to be incorrect (typically incorrect scaling on voltages), as well as lists of supported commands (including battery tests), and links to sources for that information. The DDL is also part of the NUT website, but as with the HCL, it is also stored on GitHub. Getting back to Marc's original question, this is all we have collected for Salicru: https://github.com/networkupstools/nut-ddl/tree/master/Salicru That is not to say that other models don't work, but that information was harvested from a post about USB errors, so I wouldn't consider that a ringing endorsement. At the very least, I would not want use that particular hardware for a critical system that needs monitoring. (It may be fine as a standalone unit for providing power; unfortunately, we don't get much hard information on that, since most NUT users are interested in monitoring.) > ______________________________________________ > Nut-upsuser mailing list > Nut-upsuser@alioth-lists.debian.net > https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/nut-upsuser _______________________________________________ Nut-upsuser mailing list Nut-upsuser@alioth-lists.debian.net https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/nut-upsuser
