September 23, 201411 yr Same ups? Nope. But I've never seen the issue on either system before last week, and with an average of 2 power cuts a day for the last three years, I would think that I would have noticed it before! Either the nature of our power cuts has changed, the behaviour of the UPS has changed, or the behaviour of apcupsd has changed. Have you tried reverting back to the version 3.14.10 package of acpupsd?
September 23, 201411 yr Author Same ups? Nope. But I've never seen the issue on either system before last week, and with an average of 2 power cuts a day for the last three years, I would think that I would have noticed it before! Either the nature of our power cuts has changed, the behaviour of the UPS has changed, or the behaviour of apcupsd has changed. Have you tried reverting back to the version 3.14.10 package of acpupsd? I don't see the value in that. The apcupsd only reports what the UPS is doing. Apcupsd has no control over the decision of the UPS to transfer to battery.
September 23, 201411 yr Same ups? Nope. But I've never seen the issue on either system before last week, and with an average of 2 power cuts a day for the last three years, I would think that I would have noticed it before! Either the nature of our power cuts has changed, the behaviour of the UPS has changed, or the behaviour of apcupsd has changed. Have you tried reverting back to the version 3.14.10 package of acpupsd? I don't see the value in that. The apcupsd only reports what the UPS is doing. Apcupsd has no control over the decision of the UPS to transfer to battery. Well I was looking through the release notes for 3.14.12 bug fixes " Fix issue with certain Back-UPS USB models repeatedly cycling power on/off after killpower is issued" Granted this may not apply to him but it's an example of bug that could cause an issue like this. Or a bug fix that fixes it for others and breaks it for some. And he is using Back UPS models.
September 25, 201411 yr Author A new release of apcupsd is available. I added help. Click the help button to see the help text. I also changed the default Timeout from 300 to 3000. 300 was too short and caused the UPS to force a shut down quickly. If you want a quick shutdown, then change it back. It's only a default value.
September 25, 201411 yr I had a slightly strange happening during power cut last night ... The machine appears to have shut down cleanly, with no automatic parity check on restart. However, I was observing as the system went off and the computer had not powered off at the time that the UPS shut off. I note this in the syslog, and wonder whether it would have prevented shutdown from turning the machine off: Sep 24 22:05:05 Tower rc.unRAID[25619][25647]: Running: "/etc/rc.d/rc.mpop stop" Sep 24 22:05:05 Tower crond[1515]: exit status 74 from user root /usr/local/bin/mpop -a -C /usr/local/etc/mpop/mpoprc 1> /dev/null 2>> /var/log/mail.log mpop is set as a cron job, to run every five minutes. Clearly, it had just started five seconds before powerdown asked it to stop. There is no defined interface for stopping mpop from running, so I simply issue a killall followed by a killall -9. This has, obviously, resulted in the "exit status 74" which, I think, normally signifies a time out. Perhaps I need to prevent mpop from starting once a power failure is detected, by removing the cron job entry. What is the best way of detecting a powerfail condition before powerdown is invoked?
September 30, 201411 yr Author A new version of Apcupsd is available. The major change is that the powerdown plugin is no longer required for apcupsd to work. Without the powerdown plugin, apcupsd will use the core powerdown. Because of moving plugins to dockers, the need for a complicated shutdown is now minimized. If you only use a few plugins that are well behaved (docker, cache_dirs, SNAP as examples) the optional powerdown plugin is not really needed. The powerdown plugin is only needed if you have special shutdown requirements, you want control of some additional start and stop operations with the K and S scripts, or you want to keep logs for historical purposes. This change is more in line with the LT philosophy to keep plugins to core functions and OS extensions, and not having one plugin dependent on another. Those of you that are heavily dependent on apcupsd working 100% reliably, should run a power failure test to verify that apcupsd will work properly on your system. I have limited resources to test with so I cannot cover all cases. PeterB, this is for you. Please run a power failure test to verify all is well on your system with this new version before you have a power failure.
October 1, 201411 yr PeterB, this is for you. Please run a power failure test to verify all is well on your system with this new version before you have a power failure. I'm not sure what you mean by 'a power failure test'. If you are suggesting that the power be pulled as a test, then the outcome will be little different from an apcupsd controlled powerdown (without the powerdown plugin). I am cautious of removing the powerdown plugin before a webgui 'array stop' can stop the array cleanly without further user interaction. Allowing the apcupsd-controlled shutdown to stall is little better than an abrupt power off because the ups batteries have a limited run time.
October 1, 201411 yr Author PeterB, this is for you. Please run a power failure test to verify all is well on your system with this new version before you have a power failure. I'm not sure what you mean by 'a power failure test'. If you are suggesting that the power be pulled as a test, then the outcome will be little different from an apcupsd controlled powerdown (without the powerdown plugin). I am cautious of removing the powerdown plugin before a webgui 'array stop' can stop the array cleanly without further user interaction. Allowing the apcupsd-controlled shutdown to stall is little better than an abrupt power off because the ups batteries have a limited run time. The change in the apcupsd was the powerdown script called to do the shutdown. It was /sbin/powerdown which is a specific script to the powerdown plugin that is not in the stock unRAID. I changed it to /root/powerdown which is a symlink to the powerdown script for both stock and powerdown plugin. This lets apcupsd work in a stock unRAID. I'm not suggesting that you remove the powerdown plugin. In your situation, leave it. There are a few things you can do to feel confident that the apcupsd will shut down your system properly. - Run 'powerdown -r' from the command line and verify that a clean reboot happens. - Stop the UPS daemon, set the 'Timer' field in the apcupsd webgui field to 30 and restart the UPS daemon. This will invoke a UPS shutdown in 30 seconds so it won't drain the battery down for this test. Cut power to the UPS and confirm a clean shutdown. I just turned off the power strip to my UPS when I was testing.
October 3, 201411 yr I just installed this plugin, and it seems to have installed fine; thanks. When I started the plugin it shows Daemon: Running UPS: Lost Communication Is this a concern? Should I do anything to 'fix' the Lost Communication issue? I'm running v6beta10a
October 3, 201411 yr I just installed this plugin, and it seems to have installed fine; thanks. When I started the plugin it shows Daemon: Running UPS: Lost Communication Is this a concern? Should I do anything to 'fix' the Lost Communication issue? I'm running v6beta10a Lost communication is an issue. If the unRAID cannot communicate with the UPS it has no way of knowing when it's running on battery power, so it cannot safely shut down and will turn off (ungracefully) when the battery runs out of juice. It sounds like you haven't chosen the correct cable type for your UPS perhaps.
October 3, 201411 yr Author I just installed this plugin, and it seems to have installed fine; thanks. When I started the plugin it shows Daemon: Running UPS: Lost Communication Is this a concern? Should I do anything to 'fix' the Lost Communication issue? I'm running v6beta10a Click on the help icon and you'll get a description of how to configure the settings,
October 3, 201411 yr Click on the help icon and you'll get a description of how to configure the settings, Wow! I didn't know that worked like that, very nice. Lost communication is an issue. If the unRAID cannot communicate with the UPS it has no way of knowing when it's running on battery power, so it cannot safely shut down and will turn off (ungracefully) when the battery runs out of juice. It sounds like you haven't chosen the correct cable type for your UPS perhaps. Hmmm... My UPS is connected with a USB cable. it's a Back-UPS XS 1500 I've tried changing the cable type to all the choices, except custom, because it required something to be put in a box which appeared when I selected custom. I changed the UPS type from UPS to apcsmart, and that didn't help either. I suppose it's possible, however unlikely, my UPS cable is bad, but it all worked fine back on version 5. Suggestions?
October 3, 201411 yr Author Click on the help icon and you'll get a description of how to configure the settings, Wow! I didn't know that worked like that, very nice. Lost communication is an issue. If the unRAID cannot communicate with the UPS it has no way of knowing when it's running on battery power, so it cannot safely shut down and will turn off (ungracefully) when the battery runs out of juice. It sounds like you haven't chosen the correct cable type for your UPS perhaps. Hmmm... My UPS is connected with a USB cable. it's a Back-UPS XS 1500 I've tried changing the cable type to all the choices, except custom, because it required something to be put in a box which appeared when I selected custom. I changed the UPS type from UPS to apcsmart, and that didn't help either. I suppose it's possible, however unlikely, my UPS cable is bad, but it all worked fine back on version 5. Suggestions? Generally the default settings work. You might try unplugging and re-plugging the cable. Make sure the cable is in the right port on the UPS. Try plugging into another USB port on your server. The "custom" setting enables the other field so you can enter the cable number. There is a remote chance that the V6 unRAID does not have the USB driver for your hardware.
October 3, 201411 yr Click on the help icon and you'll get a description of how to configure the settings, Wow! I didn't know that worked like that, very nice. Lost communication is an issue. If the unRAID cannot communicate with the UPS it has no way of knowing when it's running on battery power, so it cannot safely shut down and will turn off (ungracefully) when the battery runs out of juice. It sounds like you haven't chosen the correct cable type for your UPS perhaps. Hmmm... My UPS is connected with a USB cable. it's a Back-UPS XS 1500 I've tried changing the cable type to all the choices, except custom, because it required something to be put in a box which appeared when I selected custom. I changed the UPS type from UPS to apcsmart, and that didn't help either. I suppose it's possible, however unlikely, my UPS cable is bad, but it all worked fine back on version 5. Suggestions? Ah, I've got an XS 1100, and have been experiencing "communications lost" messages, but the connection does return a little while later. I have also been experiencing prompts to change the battery. For this reason, APC have been replacing my unit when it failed. The first two lasted just four weeks each, the third one has lasted around two months. Now I'm told that they have reproduced the problem with the second returned unit and tracked it down to a firmware fault. I am awaiting another replacement, with updated firmware. I have found that increasing the priority of the apcupsd process does improve matters.
October 3, 201411 yr Yeah, I'm 90% sure the default settings worked fine on version5. My USB cable is USB on one side, telephone on the other. There are only 3 places I can plug the telephone plug on the server, one says USB data port, the other 2 are network in and out. I have it in the USB data port plug. I switched the USB connection to another/different USB port on my server. I changed the UPS type to USB and apcsmart, and tested both. Still shows lost connection. I'll try to see if LT can confirm the driver is in this version. But, it's an APC brand (fairly popular?) model, so I'd be surprised if the driver is missing.
October 3, 201411 yr Ah, I've got an XS 1100, and have been experiencing "communications lost" messages, but the connection does return a little while later. I have also been experiencing prompts to change the battery. For this reason, APC have been replacing my unit when it failed. The first two lasted just four weeks each, the third one has lasted around two months. Now I'm told that they have reproduced the problem with the second returned unit and tracked it down to a firmware fault. I am awaiting another replacement, with updated firmware. I have find that increasing the priority of the apcupsd process does improve matters. Interesting. I think my unit is approaching 3 years old. I wonder if they will do anything for me under warranty. How to contact them? Also, how can I increase the priority of the process to see if that helps me here?
October 5, 201411 yr Just to be clear ... do you ever get a successful connection? In my case, I get intermittent "communications lost" messages, but the connection is re-established a few seconds later. I believe that the 'driver' being referred to above is not related to your UPS, but the Linux driver for your host USB interface. What happens if you disconnect the USB plug from your computer, and then reconnect it? You should see something like this in the log: Oct 4 18:32:18 Tower kernel: usb 2-1.1: USB disconnect, device number 3 Oct 4 18:32:21 Tower kernel: usb 2-1.1: new full-speed USB device number 4 using ehci-pci Oct 4 18:32:22 Tower kernel: hid-generic 0003:051D:0002.0004: hiddev0,hidraw2: USB HID v1.10 Device [American Power Conversion Back-UPS XS 1100CI FW:920.T2 .I USB FW:T3 ] on usb-0000:00:1d.0-1.1/input0 Oct 4 18:32:25 Tower apcupsd[1947]: Communications with UPS restored. If that doesn't work, try connecting another USB device to that host interface port.
October 5, 201411 yr ding, ding, ding we have a winner! If I plugged it into several other slots, nothing, but I kept moving it, and one of them worked fine... Oct 5 09:28:38 media kernel: usb 1-1.5: USB disconnect, device number 3 Oct 5 09:29:25 media kernel: usb 1-1.5: new full-speed USB device number 4 using ehci-pci Oct 5 09:29:25 media kernel: hid-generic 0003:051D:0002.0002: hiddev0,hidraw0: USB HID v1.00 Device [American Power Conversion Back-UPS BX1500G FW:866.L4 .D USB FW:L4 ] on usb-0000:00:1a.0-1.5/input0 Oct 5 09:29:28 media apcupsd[1869]: Communications with UPS restored. I suspect this is a result of my efforts to get windows 7 passthru working, which I had to abandon several months ago. I guess it's time to revisit that work, and 'fix' my issues there. Thanks for the great lead. My UPS is showing everything fine now.
October 11, 201411 yr Author There is an updated Apcupsd plugin available. Changes were to the plugin. No functional changes were made. The emhttp files (apcupsd.php, apcupsdctl.php, apcupsd.page, apcupsd.png, and READ.me) were moved to a downloaded tar. This is more in line with the LT strategy for plugins.
October 12, 201411 yr Author An updated apcupsd plugin is available. I made a few changes to make the plugin easier to read and maintain.. The icon was changed to one with a transparent background to be compatible with the Dynamix black theme.
October 12, 201411 yr ...to be compatible with the Dynamix black theme. Say what?!?!? Is Dynamix available now? I'll have to go take a look. I've been hoping/waiting for that to happen, but didn't see anything. ** Not for v6 it seems
October 12, 201411 yr Author ...to be compatible with the Dynamix black theme. Say what?!?!? Is Dynamix available now? I'll have to go take a look. I've been hoping/waiting for that to happen, but didn't see anything. ** Not for v6 it seems I'm only getting ready for it. I know how it works from my V5 experience.
October 12, 201411 yr Ah, makes sense. Probably good, I have enough other things I need to 'fix' before adding that to the mix
October 12, 201411 yr Ah, makes sense. Probably good, I have enough other things I need to 'fix' before adding that to the mix +10000! I read this and thought you were nuts! lol
Archived
This topic is now archived and is closed to further replies.