Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[ojosch]

I eagerly anticipated the arrival of this DNS-323 along with a new APC Back-UPS ES 750 for a nice xmas present for my mother this year. Well guess what? I plug it all in and in the 'Status' page of the 323 it shows the UPS, so I pull the plug and monitor both the percentage and 'On Battery' status for the entire life of the drain test, and it went right down to 1% and I also had my laptop plugged into the UPS so that when the 323 would die I could view the power indicator on my laptop to see if it reports being on battery or AC power. Well, at the exact same moment when the power on my laptop switched from AC to battery (indicating UPS power gone), the DNS-323 died right then too.

Shouldn't this thing shutoff like around 3-5% or something? Mine is definitely NOT doing what the FW 1.06 release notes state that it should be doing. Could somebody please tell me how to get this working? Like I said above, I can see the percentage on the 323's 'Status' page just fine,  and at one time when the web GUI timed out due to inactivity and I logged back in, it gave a message saying to save your work because the 323 will be shutting down soon, but when it got to 1% I was watching closely and... nodda

Any ideas? I specifically got this setup for my mother since her last RAID 1 NAS setup suffered data loss from a power outage. So I thought this setup was the PERFECT CURE, but I kinda feel a little betrayed if this is what they call safe shutdown. Even the battery-low warning message of the web GUI is written in poor English and has no space between 2 sentences, and poor translation. This bit of tackiness leads me to believe that the rest of the FW package may be poorly written and not proof-read for errors to a very thorough extent. Is this all we can expect from this seemingly fine product from Dlink? 

[ojosch]

Well, with a second tiny strand of hope, I tried it again, and nope... nodda.

This time I got the battery to around 40% percent and just to be safe, I logged OUT of the web GUI in case any active sessions might disable auto shutdown, and then just waited and watched. When the UPS went from 4 beeps per 30secs down to constant beeping, I watched very close, and got the same results. Right when the laptop switched to battery, the DNS-323 died out. And it sounded like the unsafe way because when it died the hard drives made the clunk sound. I am loosing faith in this thing fast. Is there something that needs to be enabled in the UPS settings? I haven't installed the software disc into my Windows box yet so I don't know if there are any configurations that can be done to the UPS or not. But that's the only thing I can think of is that I need to turn on something in the UPS.

Also, it's too bad that the FW 1.06 doesn't have any more adjustment to the UPS settings, like maybe 'Enable' or 'Disable', or maybe 'percentage of battery remaining before initiation of shutdown sequence' or something like that. It seems awfully 'Mac' to just not have any more control over these settings, other than if it's plugged in it's supposed to just work, and if not, no. And at least if it were Mac, I have no doubt that it would just work

[ojosch]

Ok, so tonight, I hooked up a different UPS, an APC Back-UPS ES 725 with a different USB sync cable that I bought on Amazon the other day (the p/n that is supposed to ship with this UPS), and I also updated the firmware just for kicks from 1.07 (latest release) to 1.08 (latest Beta build). Then I re-ran the same tests as before, and it still fails exactly as before. These are the same UPSs that Dlinks recommends, and as far as I know everything is all set up right, so why am I having this issue? I wish I could get an engineer for this team who designed this feature to explain to me the correct test case to make this work as it should. I don't get it. It makes me think that maybe everyone else who set this up maybe never tested it either and maybe this just doesn't work at all. My hardware is version B1. Could this be an issue? Has anyone tested this setup with B1 hardware? I'm not sure what to do now. If any services are running can this cancel out the shutdown sequence? The only service I am running is ftp. If anybody has any suggestions for me I'm all ears. Thanks

[ojosch]

If I go to 'System' and click 'Shutdown' from there it does manually shutdown. So it CAN shutdown, it just doesn't seem to ever run the script to shut it down as you watch the status percentage power left area in the Dlink web GUI go down right past 1% and lower til the UPS pukes out, and the 323 just pukes right along with it. Nobody has had this problem?

[ojosch]

Is there anybody here who knows how nut works who might have an idea on where the problem could lie? I guess I am going to install funplug and try to go in there and figure out the nuts and bolts of it.

[ojosch]

Ok, I got it working now. I had to install fun_plug, then install the nut package, and then made a fun_plug start-up script that copies 2 files: upsmon (from /ffp/sbin) to the ramdrive space (/sbin/), and upsmon.conf (from /ffp/etc/) to the ramdrive space (/etc/), then it runs from there. Then I just made the upsmon.conf file point to the right driver that the DNS-323 already loads out of the box, and upsmon DOES now finally shutdown unit.

But upon later research, I found that the shutdown command was causing an unsafe shutdown, so I installed the cleanboot fun_plug which now stops all of the processes that run off of the HDDs, then umounts the drives and then proceeds to shutdown the unit safely. Before I did this, the stock process that upsmon runs is "shutdown -h +0" which is invalid in the dlink firmware, so I had originally changed it to /sbin/poweroff which seemed to halt, but not power down unit. And it wasn't safe halting. So cleanboot installs new binaries called "shutdown", "halt", and "reboot" which do safe shutdown.

So now that this is working, I have now found that dmesg is telling me I need to run e2fsck to check filesystem of the HDDs, so to do this, I need to umount the /dev/md0 drive, so I have modified my telnet start-up script to copy it's files to the ramdrive first and running the daemon from there, so it is not dependent on the HDD, and I've shut off any other daemons that were running off the HDDs, but for some reason, I still CANNOT run umount -f /dev/md0. I just get the error: "Cannot forcibly unmount /dev/md0, Device or resource busy".

So if I can figure out how to run fsck'r, and get the errors to go away, then I have a nice system here, which is capable of shutting itself down after it depletes a UPS. It shuts down on "LOW BATT" which happens around 10% UPS battery power.

I'm almost tempted to just install the new beta version which supposedly has filesystem checker built-in, and I've made all of the 3 fun_plugs I use run off of the ramdisk space, and nothing is dependent on the HDDs so maybe the beta 1.08 can figure out how to unmount that HDD, since I can't. I'm still a Linux noob I guess.

By the way, this string has seemed like a pretty one sided conversation to me. People are pretty talkative here 

[gunrunnerjohn]

By the way, this string has seemed like a pretty one sided conversation to me. People are pretty talkative here

I suspect not many folks are tinkering with the UPS, which is probably why they don't have much to add.

[iuliand]

Nice work!
But I really hope dlink will fix this issue for the rest of us too...