David S. Madole
2009-Aug-27 15:04 UTC
[Nut-upsuser] Serial comm errors with bcmxcp on HP R3000 XR
I have setup nut-2.4.1 with the bcmxcp driver for my HP R3000 XR unit and although it appears to work, I receive the following sorts of errors in syslog: Aug 27 10:33:05 sv18 bcmxcp[6373]: Communications with UPS lost: Receive error (data): got 0 bytes instead of 64!!! Aug 27 10:33:15 sv18 bcmxcp[6373]: Communications with UPS lost: Receive error (data): got 0 bytes instead of 64!!! Aug 27 10:38:57 sv18 bcmxcp[6373]: Communications with UPS lost: Receive error (data): got 0 bytes instead of 64!!! Aug 27 10:40:04 sv18 bcmxcp[6373]: Communications with UPS lost: Receive error (Request command): 0!!! Aug 27 10:43:38 sv18 bcmxcp[6373]: Communications with UPS lost: Receive error (Request command): 0!!! After about 15 hours the driver exits and has to be restarted. Here is my ups.conf: [ups2] driver = bcmxcp port = /dev/ttyS1 desc = "HP R3000 XR" baud_rate = 19200 The system is running Slackware Linux 12.1 and nut was built from source. Any ideas? Thanks, David -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.alioth.debian.org/pipermail/nut-upsuser/attachments/20090827/1b13077c/attachment.htm>
Kjell Claesson
2009-Aug-27 17:01 UTC
[Nut-upsuser] Serial comm errors with bcmxcp on HP R3000 XR
Hi David,> I have setup nut-2.4.1 with the bcmxcp driver for my HP R3000 XR unit > and although it appears to work, I receive the following sorts of errors > in syslog: > > Aug 27 10:33:05 sv18 bcmxcp[6373]: Communications with UPS lost: Receive > error (data): got 0 bytes instead of 64!!! > Aug 27 10:33:15 sv18 bcmxcp[6373]: Communications with UPS lost: Receive > error (data): got 0 bytes instead of 64!!! > Aug 27 10:38:57 sv18 bcmxcp[6373]: Communications with UPS lost: Receive > error (data): got 0 bytes instead of 64!!! > Aug 27 10:40:04 sv18 bcmxcp[6373]: Communications with UPS lost: Receive > error (Request command): 0!!! > Aug 27 10:43:38 sv18 bcmxcp[6373]: Communications with UPS lost: Receive > error (Request command): 0!!! >This is communication errors. The protocol is a binary protocol with multi packet, and header and checksum. And instead of guessing that packets and information is right, the communication is tested for any fault condition. So the first error: The header say that the packet should contain 64 byte, but it return 0. The second say that you have sent a request command but the blocknumber returned is not the same as the command.> After about 15 hours the driver exits and has to be restarted. >To many errors maybe.> Here is my ups.conf: > > [ups2] > driver = bcmxcp > port = /dev/ttyS1 > desc = "HP R3000 XR" > baud_rate = 19200 >OK.> The system is running Slackware Linux 12.1 and nut was built from source. > > Any ideas?So check your cabling. Use hi grade serial cables. Look out for ground loops in the equipment. For my information you may run this command after halting the upsmon and upsd and upsdrv. /path/to/bcmxcp -DD -a ups2 -u root Kill it by ctrl-c after it start repeating the outlet number. This make the driver spit out lot of debug info. Copy the info and mail it to me. Regards Kjell
Kjell Claesson
2009-Aug-28 16:07 UTC
[Nut-upsuser] Serial comm errors with bcmxcp on HP R3000 XR
OK David,> Kjell, > > Thanks for your time in replying. > > The cable is a regular commercial shielded cable and is only six feet > long. Grounding should not be an issue, everything is on a single > circuit with a single ground and is powered through the UPS itself, all > in one cabinet, so all grounds should be at the same potential. >That is good. Should not be any problems. Then the problem may be the polling frequency. The ups is sending a lot of data for each request. But as you are running it at 19200 it should be OK. I have my PW5125 on 9600 bps and have no problem. What is your system spec ? Linux freebsd ? Kernel version ? I think there was some glitch around version 2.6.25 of the kernel that make the serial port behave strange. Can you relate the errors in the log to any activity? Is the time random or is it spread evenly over time ?> Attached is the debug information you requested.Thank you for the info. It looks OK. No errors that info is missing. Older firmware may have some problems, but it look fine on yours. Regards Kjell