I have a ReadyNAS Duo attached to a Netgear GS605 switch and then to a TP-Link router (it was connec ted to a Netgear WNR2000 but I swapped them in an attempt to solve this problem, which is unchanged - I also swapped the switch from a GS605 so we can rule that out).
The router assigns a reserved IP address and my Win 7 PC can see the NAS using the IP address or text hostage. RAIDar can see it and I can access the admin page in a browser. Everything works for 15-20 minutes, then the NAS vanishes - sometimes all at once, other times bit by bit - you can access it via IP but not hostname; then Windows won't see it at all, but the admin page still works; then even RAIDar can't find it - although it does still seem to be talking to the switch.
At that point all you can do is pull the plug and restart - whereupon everything works... for 15-20 minutes. The logs only show evidence of the plug being pulled - there's no indication of any problems prior to that.
Having searched around here I found references to similar problems caused by jumbo frames. Turned them off and it happened again during a backup. Not impressed...
I'm assuming it's the original Duo and not the new Duo v2. What version of firmware is on the NAS? (v1 Duo's run 4.1.x and new v2's run 5.x.x).
If it's the original Duo, older versions (4.1.6 and lower) have a bug with media scanning that was corrected in later versions. See here for details: http://home.bott.ca/webserver/?p=260
So, if not running current version (4.1.8), please update to latest version. There's also a 4.1.9 beta in the pipeline - viewtopic.php?f=17&t=59222.
If you're already running the latest version, it could be related to disk problems and you will need to rule them out as the source. Have a look here for things to look for and/or try: http://home.bott.ca/webserver/?p=388
These two are the most critical for drive 1: Current Pending Sector 51 Offline Uncorrectable 49
If you don't have a backup, you should make one. If you see any signs at all that these two parameters are changing (or that Reallocated Sector Count is growing), then you should replace it without delay. Personally I would replace this drive now. As sphardy indicated, it could explain the vanishing act.
These two are the most critical for drive 1: Current Pending Sector 51 Offline Uncorrectable 49
If you don't have a backup, you should make one. If you see any signs at all that these two parameters are changing (or that Reallocated Sector Count is growing), then you should replace it without delay. Personally I would replace this drive now. As sphardy indicated, it could explain the vanishing act.
Thanks. I've pulled it for now and the NAS has stayed up, so it does seem to have been the problem. I'm not sure having a NAS randomly disappearing from the network without any clue as to the cause is the most robust way for a RAID drive to handle drive errors...
Disks can and do fail at any time and in a large variety of ways. Unfortunately sometimes this can't be handled gracefully.
It is important to remember that RAID is not a backup. If you store important data primarily on the ReadyNAS you should backup that data to a USB disk, another NAS or some place else. See Preventing Catastrophic Data Loss
