Unfortunately no one can be told what fun_plug is - you have to see it for yourself.
You are not logged in.
So one day 1, when I was transferring data from the DNS-323 over to the DNS-343, I experienced the freeze... you know, it won't do anything, I tried to shut it down and it didn't shut down... when I power up, it gives me a sync time of 1000min... anyway, long story short, I had RMA'd a new one, the new one, I noticed that I would get intermittent slow downs... ie. I'm reading from the DNS-343 at 8-12MByte/sec on a 100MB connection, then it'll slow down to 100k/s and be on/off for 5-10min, then go back to normal and transfer at 8-12MByte/sec...
Here's my system info:
DNS-343 using Raid 5, ext3, 4x1.5TB WD Green drives, all drives have TLER on, and head parking off.
FW is 1.03, drives formatted with this fw.
all other services off.
static ip, SMB share, connected to Linksys WRT54GL which is then connected to laptop
I tried downgrading FW to 1.02beta, which didn't help.
I just recently installed funplug and when it happens, I notice when running top that the IO CPU percentage goes to 99%, and load is as high as 2.85... that's probably what's causing the slow downs... but the question is, why is IO going to 99% ?? There doesn't seem to be anything else via top that is out of the ordinary (ie. memory or bad processes).
Has anyone experienced this or have any suggestions ?
I called DLINK tech support and they opened a case file, but said they would need to replicate it, so I don't have much faith in them.. they didn't even ask me for screen captures or any information...
I'm really frustrated... I spent $400 on this thing, plus $600 on HD and it won't work!!!
Offline
here's a screenshot of my torrents...
Offline
is no one else experiencing this problem ?
Offline
I have a couple machines that easily transfer ISO's of DVD's and such with no problem.
I've found of the 2, the network card makes a big different as one is faster (AMD quad core) but has a realtek NIC that transfers at about 100Mb/s (w/9K jumbo frames) vs a Pentium4 3GHz HT that with it's Intel Pro/1000MT NIC (and jumbo frames) transfers at about 320Mb/s. (yes, those are both in BITs)
12.5MB (Bytes) is about the max for a 100Mb/s connection. I'm actually surprised you get that after ethernet frame overhead.
But anyway -- the stalls are strange. I never see those.
-Ben
Offline
thanks for the post... DLINK doesn't know why it's hapening either... i'm in the process of moving all my files off the NAS and reformatting the drives...
Offline
Consider trying a direct connection with static IP's. Take any network switches out of the equation.
I've known switches of all kinds to have the potential to be flaky.
-Ben
Offline
I've formatted the drives and seems to still happen, so I'm running an extended SMART test on the drives...
I'll try the direct connection as you suggest later... I seemed to have slow down issues with both my Linksys WRT310N and my WRT54GL so I'd think it'll be low probability that both my routers/switches are bad... but it can't hurt...
Offline
after more testing, it seems like this slowdown happens consistently only on 1 of my 4 drives... even with that drive by itself in the DNS-343, I can reproduce the problem.
The drive works fine in a computer, and passes all the SMART testing and any HD test program I can throw at it... nevertheless, I am RMA'ing the drive and see if a replacement will fix this problem...
weird
Offline
Weird.
What kind of drives were they? I'm using Seagates in my system. (250GB's)
-Ben
Offline
They are WD WD15EADS drives... I have received the replacement drive and now attempting to recreate my RAID5 volume... will update once successful
Offline
I'm having a similar problem, I've assumed up till now it was due to the network in some way. I have a DNS 343 running with 4 of those drives (1.5gb cav green)
I get about 1 min of great transfer time, then it goes to 100Kps! and up for a few seconds, then back down to nothing.. my network graph is a series of short spikes from then on.
I'm going to rule the network out of the equation first, but i thought i'd better ring in, just in case.
Offline
Hi. I have a similar situation here: same 1.5TB drives (WD15EADS-00P8B0 WD Caviar Green) in a DNS-323 RAID 1 config. Static IP and original v1.07 firmware (no funplug). Auto-speed to a gigabit switch (should I use 1000 instead?).
Very slow! And seems to burst transfer data, not a regular flow. I have 6 other DNS-323 with 1TB Seagate drives on the same switch; never experienced this before.
I just enabled Jumbo Frame and set it to MTU = 9000. It was disabled before. Will report.
Let's keep posted on this please. Fingers crossed. Iam Nemo
Offline
I just enabled Jumbo Frame and set it to MTU = 9000 and it seems to be MUCH better.
Offline
Well finally the improvement was short term. My Netgear unmanaged 8-port switches do support jumbo frames (but at what MTU I don't know) and so does my main PC but it has never been implemented in the Nvidia driver (!). So I'm back at mid-level transfer speeds, not too bad I guess.
Offline
Once everything is working steadily, I usually set the DNS-323 or my DNS-343 to a fixed speed instead of auto negotiation... if you are still experiencing spikey transfer rate, then it could be the problem I had, which turned out to be a drive issue (despite it passing all tests that had thrown at it). The data transfer rate should be relatively consistent...
Offline
twg wrote:
Once everything is working steadily, I usually set the DNS-323 or my DNS-343 to a fixed speed instead of auto negotiation... if you are still experiencing spikey transfer rate, then it could be the problem I had, which turned out to be a drive issue (despite it passing all tests that had thrown at it). The data transfer rate should be relatively consistent...
Thanks, I'll keep that in mind.
Offline