Unfortunately no one can be told what fun_plug is - you have to see it for yourself.
You are not logged in.
Hi,
I've been running a DNS-232 for over a year now, and decided last week to get another device to use for media storage.
I currently have a device configured as a 500GB mirrored RAID, and got another system with two 1TB drives in a striped RAID to use as media storage.
So, got the system and disks today, plugged it all in, selected the striping option and sat back for 20-30 mins while the disks formatted. Got a message to say that all was done, and the system will reboot, and hey presto, same configuration screen _again_ asking to format the disks.....WTF ?
Selecting 'Skip' verifies that the NAS still thinks there are no disks installed...
The box only has one basic job, to recognise disks and format them correctly..., it can't be that difficult can it ?
Anyone else had this 'feature'...?
Hardware Revision B (my first box is an A) Firmware 1.05
2 x Samsung SpinPoint F1 HD103UJ 1TB SATAII 32MB Cache
Last edited by tescophil (2009-03-17 23:27:52)
Offline
Disk management is not the DNS-323 strongest suit, to put it nicely...
Offline
OK..., so what now...?
Tried different configs JOB / RAID etc and the disks will still not format.
Left it running overnight and this morning the screen was stuck at the magic '94% completed' after 12 hours...
Not funny any more.
Oh Yes, and now my right drive led is Purple..., searched for what this means on the forum and nothing actually useful came up (if you are a standard, non funplug user) apart from bad'ish...?
So, do I just pack the lot up and send it back to the supplier ?
Last edited by tescophil (2009-03-18 08:37:50)
Offline
Are you removing the partition tables between trying different configurations?
I believe the D-Link setup scripts expect the drives to have no partitions on them.
Offline
Why don't you fry firmware 1.06 ?
RAP
Offline
rap wrote:
Why don't you fry firmware 1.06 ?
That is a very good question. (I think it was directed to me, but re-reading the
thread it could be directed to tescophil)
I haven't found a compelling reason, for me, to upgrade. I have worked around
all the issues I have with FWv1.03 using custom scripts and fonz_fun_plug . This
DNS-323 is "in production" as our home data server, so downtime due to
software changes is a real big issue for my family.
I do have a second DNS-323, running FWv1.06 and FFP v0.5, that I use as a
"development" box to "try sutff out" and determine if there is something in the
new firmware (or more likely, FFP) I need to implement.
I really only use my DNS-323 as a file server - K.I.S.S. (Keep It Short & Simple).
Great piece of hardware, but really buggy D-Links scripts, but the underlying Linux
packages are sound. So, I try not to use the D-Link web GUI, and rely on fonz_fun_plug
and my scripts to get the DNS-323 working the way I want.
Last edited by mig (2009-03-18 17:24:09)
Offline
I was resistant to trying 1.06, because I tried it on my first 323 and I got the 'Disks always on' feature, so downgraded back to 1.05
But, I upgraded the new system to 1.06 and formatted as a 2TB JBOD Volume.
The format was sucessful, and the purple LED went away after the 'reboot after format' cycle.
However, it's still very poor as a user experience.
When copying my media data to the new drive, I see both drive lights active, with the right bay less active than the left, but active all the same.
Does anyone know how the data is arranged on the disks in a JBOD configuration ? I just assumed it would use one disk untill it was full, then move onto the next..., it just dosn't look like this if the drive access lights are anything to go by...
Cheers.
Phil.
Last edited by tescophil (2009-03-19 00:52:48)
Offline
tescophil wrote:
I was resistant to trying 1.06, because I tried it on my first 323 and I got the 'Disks always on' feature, so downgraded back to 1.05
But, I upgraded the new system to 1.06 and formatted as a 2TB JBOD Volume.
The format was sucessful, and the purple LED went away after the 'reboot after format' cycle.
However, it's still very poor as a user experience.
When copying my media data to the new drive, I see both drive lights active, with the right bay less active than the left, but active all the same.
Does anyone know how the data is arranged on the disks in a JBOD configuration ? I just assumed it would use one disk untill it was full, then move onto the next..., it just dosn't look like this if the drive access lights are anything to go by...
Cheers.
Phil.
This is exactly what it does - and you can verify it by pulling the second disk (hot unplug), which should be the left side. Assuming that you have not filled the disk, the data that you wrote should remain available - up to the first reboot (which is why you need to hot unplug the disk for this test), after which you will lose access to it.
Over on the Dlink forum I said the data was lost, and a mod got a little peeved - so to be clear, the data is still on the disk, you have not lost it, but you can no longer access it without a lot of grief .
Think twice about using JBOD - lost data or inaccessible data (same difference to me) - it's your data at risk
Offline
Cool...,
I have two devices, one in a RAID 1 config for better reliability, which I also backup to another drive, and the larger new array for stuff that I can easly replace...
I didn't realise that if one of the disks fails in a JBOD array, then its stuffs the other one as well... what's the point...?, may as well stripe the disks instead and get better performance ?
Cheers.
Phil.
Offline
Stripe the disks and get better performance, eh?
Now I could be wrong here, but I have a strong feeling that the only time you're going to see better performance with striped disks (ie RAID0) is when those disks are directly attached. RAID0 increases performance only when the disk itself is the bottleneck, so when you attach them via the network, as in a NAS, the results might not be what you expect - in this particular case, RAID0 perfomance is no better (or worse) than RAID1, JBOD or standard volumes.
From my point of view, there is no reason to use JBOD or RAID0 on this device - increased risk with no reward - RAID1 or standard volumes.
Offline
The only reason, in my opinion, to use RAID0 is if you are manipulation single files larger than a single disk. If you are doing that, you may want to rethink the hardware you're using anyway.
If you are interested in how the unit works, and not having issues with the DNS's scripts, search for upgrding 1.03 to 1.04. I wrote a cript a while back that you can look at. It step by step formats and creates a RAID1 array, and populates the config files for the DNS. I tried to remark most of it, and I tried to make it modular, so you can easily see what's going on. (Don't run it unless you are actually doing the upgrade.)
Offline
I found the same problem when installing 2x 1.5TB disks in this unit, Rev B w/FW 1.6, some google-fu quickly shows you that to get a reliable result you should install 1 disk at a time in normal mode, and then reconfigure them after the unit has both disks in that mode and working. I can't quite remember if this is an issue with disks of 1TB or greater but IIRC it was only an issue with drives this large.
It doesn't combat the point however that this unit's firmware is brain dead when it comes to the basics but for the price, and thanks to fun_plug capability you definitely can't find a better device for the price point.
In response to some of the other posters on the forum regarding disk speed, I can guarantee that you will not see better performance out of this device in a Raid 0 configuration, I am running a Raid 1 and have verified the the limitation for IO on this unit is the network connection (with raid1/0). however, I am not sure how you would fare in JBOD mode, a single disk may not be quick enough to keep up with the IO speed of the network. If you are aiming for performance be sure that you have a machine that can do 15k jumbo frames to talk to this device, it becomes quickly CPU bound when doing large transfers and causes it to choke up a bit.
as a measure of safety, because i haven't come to fully trust the raid in this machine, I do a nightly rsync of critical data to my windows server (running delta copy) just to be sure I have a place where things like photos are definitely safe!
Offline
Actually the network connection is not the limitation - it's good for something in the vicinity of 400 mbit/sec or approximately 50MByte/sec, well over the maximum system throughput that I have measured at approximately 30MByte/sec (read).
Of course you won't see this unless you specifically test for it, using a test utility that specifically measures network throughput.
Offline
bruor wrote:
as a measure of safety, because i haven't come to fully trust the raid in this machine, I do a nightly rsync of critical data to my windows server (running delta copy) just to be sure I have a place where things like photos are definitely safe!
Even if you do come to fully trust the raid in this machine, you should keep that nightly rsync going. Raid won't protect you from accidentally deleting your photo collection, and this thing doesn't have a network recycle bin.
Offline