Unfortunately no one can be told what fun_plug is - you have to see it for yourself.
You are not logged in.
This morning I tried to create a backup tar of a directory and I was informed that the drive was full. It's possible that it was (although I was surprised as I thought I had a couple of hundred gig spare of my 1TB mirrored volume).
I then using WinSCP, instructed the device to delete about 20 very large files (old backups). After waiting a few minutes, I decided that it would take forever through WinSCP and deleted them through ssh.
However, from the time that I got the 'drive full' error, I have not been able to access the drive through my Windows Explorer drive mapping and the drive lights have both turned pink. Other than that, it appears to be working.
I can still see the files stored on it, via ssh and WinSCP and the webserver still seems to be serving files (even those requiring MYSQL access on the machine).
So why are my drive lights both lit steady pink? I checked the manual and it mentions that orange means the drives have failed, but it doesn't mention pink, which these very definitely are!
Any advise please>
Thanks!
Offline
Restarted the device and back to blue again now. Odd! Very odd.
All ok now though.
Offline
I've never gotten a 100% solid explanation of what causes the pink/orange lights to go on. (I think pink and orange are the same - people have different opinions on the colors, and the LED's may vary.)
In my experience the most common explanation is that the root filesystem has filled up. Note that this is the root filesystem, and not your storage disks /mnt/HD_a2, etc. The root filesystem is fairly small and is in RAM.
The root filesystem can be easily filled up if you have hacked the device to run telnet or ssh and you are running scripts or doing something that uses space in this filesystem. For example, copying a file to /tmp uses up space in this filesystem. So if you run a script to make zip files, backups, or whatever and that script uses /tmp for temporary files, you could easily fill up the filesystem and cause the lights to go on. Most people write their scripts carefully to make sure temporary files go elsewhere.
In my case, I run a lot of scripts from cron, etc and it took some effort to make sure I didn't fill up the root filesystem. In particular, I use rsnapshot to archive /dev/HD_a2 to /dev/HD_b2. I notice that if I have a LOT of updates to the a2 disk, the subsequent rsnapshot can fail and cause the pink lights to go on. Rebooting does fix it. Since the rsnapshot failed, it gets rolled back and will sometimes fail on the next attempts. What's weird is that it will eventually do a good snapshot, which I have not figured out.
So, if this happens again, try to ssh into your machine and run the df command to see if rootfs is 100%. On mine, it is usually 80%.
I think what happened to you is that your rootfs filled up. You deleted some files on your data disk, which really didn't help anything. When you rebooted, the RAM rootfs got reset and everything looked OK. I think what started all this was making your backup tar of a directory, as you said in the first post. Your tar file went into the rootfs and filled it up.
Offline
Pink & orange are definitely NOT the same - the LED is designed to display two colors, blue & orange, however, it can apparently display both simultaneously and depending on the intensity of the two colors, the end result can appear to be a pinkish/purplish/white.
The orange light is defined as a drive failure and can be forced, for example by pulling one drive out of a functional RAID array.
Offline
Sorry, you're right. The discussions I saw were about pink vs purple. I've never seen an orange light, which is a good thing!
Offline
i used to get the pink lights when I had PNP enabled with FW 1.06. I dont think I ever had a need to enable again but I would assume that has been resolved
Offline
I also get the ping lights on both disks and also get an email alert saying both disks have failed.
This always occurs when FTPing large amounts of data to the device on one disk only. I am not touching the other disk, but both go to pink. The large amount of data can be over 1GB (say 4 files of 200-300 MB each).
I can usually get into the web interface and reboot the device and it gets back to normal. Note, I have no hacks what-so-ever applied to this unit. It is running 1.08 software.
Has anyone else this experience?
Offline
Yeah, I use Macrium Reflect to back up my PC Windows XP partitions. For awhile, I tried storing the resulting (large) image files on the DNS-323 via samba and would frequently see that type of problem. I gave up and now save the Macrium images on a USB drive attached to the PC. Too bad, since this is the type of thing that I really wanted to store on the 323. Now, I rsync selected parts of the Win XP file tree to the DNS-323, in addition to the periodic partition backups.
My guess is that samba is failing on the 323 under the load, since during the few times I looked at it, it seemed that the samba related processes were not running. I also guess that there is a watchdog process on the 323 that checks to see if the samba processes are running and then turn the lights on if not. I HAVE been able to recover from the situation by simply restarting the samba services. I have not investigated further.
One thing to try is to install a "better" samba; I think that there is a howto on the wiki for that. I may give it a shot someday.
Also, since you are running a non-hacked 323, it might be worth looking at the D-Link forum to see if there are any clues there, since most of those people are running the stock 323.
That's all I know; I don't know if it helps or not.
Offline