After many years of working well, other than a bad PS, I began having connectivity issues. No errors or messages on the NAS end, but the NAS would drop off the network. Would have to reboot to get it back online.
It was suggested the OS might have gotten corrupted and a rebuild might solve the issue. Well, a long time later, I finally did the rebuild. The server ran for two weeks without a hiccup, and then the connectivity issue came back. Rebooted, ran a day, rebooted, ran 4 hours.
If it isn't the OS, then can I assume hardware? I carefully blew out all of the dust when I had the server down to upgrade the drives in it and upgrade to the latest OS. If it is hardware, I am guessing we are talking about a mobo and a mobo might not be easy to get. Thoughts, ideas, suggestions?
When you say it "dropped from the network", can RAIDar still see it? If so, try leaving an SSH shell open and logged in and see if it stays connected. Then use TOP to see if some process is eating all the CPU time when it disappears. I have seen some faults drive the "readynasd" process to nearly 100%.
Hello, if I knew about RAIDar, I forgot. I just installed it to see if it can continue to see the NAS when neither my PC nor my FireTV box can access the NAS. Unless it "talks" with a different protocol, i.e. not http or? then I don't know if it will continue to see the NAS or not. I wasn't able to make an SSH connection just now. Guess I did something wrong. I didn't have the NAS up, so this is a fresh start from a cold box. Will see how long it lasts.
... Unless it "talks" with a different protocol ...
It does. It discovers the NAS (and determines it's status) using broadcast ethernet packets. This can connect with the NAS even if the IP configuration iin the NAS s incompatible with your subnet.
OK...as mentioned earlier, I started up the NAS after a few days of it being turned off. As usual, the system booted up, my shares were available, and life was wonderful, for the moment. As suggested, I launched RAIDar. I left the PC running RAIDar on. This morning, as usual lately, the PC showed the shared drives in Explorer missing and the network section of Explorer showed the NAS absent and was unable to log back in.
RAIDar show it's "healthy" status still, the page had not changed since the night before, i.e., although it was disconnected from the NAS, it didn't know it and didnt' how it. When trying to click the "admin page" option, I got the "Your NETGEAR stoarge is not accessible from this computer. Check your network settings." Browse and Locate give the same message. Download Logs, Diagnostics, and Restart all ask for login info, but fail as they too can't reach the NAS.
As far as a process causing the box to become unresponsive isn't likely. The server ran fine on the old version of the OS and never hiccupped until this problem began. The system having been reset and the new OS installed indicates the problem moved from the OLD to the new OS, indicating to me anyway, it isn't OS related. I don't have any installed apps or anything that didn't come in the default build. I am only using the box to serve files, and wasn't even doing that when it became disconnected.
I have two Ethernet cables connected at the moment, but didn't do any bonding. I can disconnect one, but don't know if it will change anything. Before I reset the server, I did have them bonded so whether they are bonded or not doesn't seem to matter. It doesn't appear to be software which means it is hardware, and if that is the case, the logic is on the mobo? And to replace the Ethernet ports, that means a mobo swap? Presumably the mobos have been long since obsoleted, and on the used market, probably the part most sought after?
update....I mentioned earlier that the RAIDar window was open and showed the server to be fine. I closed and re-launched RAIDar and it was able to see the server, and still shows it "healthy". So that means the OS is still operational, there is some Ethernet connectivity for it to still see the NAS, but then why am I losing HTTP/and or whatever other protocols (SMB) I need?
I checked the firewall. Shows the server connecting with the PC every 2 minutes using UDP. Randomly there are ICMP outgoing connections from the PC to the NAS.
RAIDar shows the NAS at 192.168.0.13 and 21. The firewall log showed all traffic on .13.
Unrelated to the NAS issues, I did add an Arlo and a Connect Home Pro. I saw no correlation to their install and this issue. I think I had this issue before either of those were installed but I would have to check dates to be sure.
All of the devices on the network are connected to a common Netgear ProSafe Plus JGS516PE Ethernet Switch.
The OS consists of many pieces -- some standard Linux and some custom. Not everything is updated with a new version of the OS. So that the problem survivied an OS update doesn't really indicate what you seem to think it does. I still believe that leaving an SSH session open and seeing if it also disconnects or, if it doesn't, what processes are running and if any are requiring high CPU use after the SMB disconnect, could be helpful in diagnosis.
but then why am I losing HTTP/and or whatever other protocols (SMB) I need?
Using two unbonded NICs connected to the same network might be at least part of the problem here. I'd try disconnecting one of the ethernet cables, and see if the connectivity improves.
Mr. Shark, Did you see that I did a full factory reset, not just an OS update on top of what was there? If I understand the process, a new image updates the embedded firmware and when new disks and a factory reset was done, there should be a fresh OS image with nothing from previously?
Tried connecting to the box with Putty, WINSCP and "Shell in a Box" and none of them is able to connect via SSH.
Never tried to SSH to the NAS so I don't know if there is something special I am supposed to be doing.
I can certainly disconnect one as I am not in need, at this point, in high data flow since I have put all of my data back on the NAS. Would the other option be to bond it?
I did find it odd that .13 is connected to eth1 vs. the NAS defaulting to eth0, .21.
That didn't go so well. Initially the server admin screen operated on .13 which was Eth1. When I bonded them, the bond IP changed to .21, which was OK, but performance dropped to nil. It too 3 minutes just to bring up the admin screen. I deleted the bond and "control" went back to .13, BUT, Eth0 is now grayed out. I went into it's settings and nothing there about enabling the port...wasn't able to do anything to enable it. Haven't tried rebooting the box as we are now at one day, 15 hours of uptime. Since nothing was changed during this troubleshooting, I expect the problem still remains and it will be hours or days before it becomes unresponsive.
I went into the NAS and enabled Round Robin, not knowing I needed to do more. Either way, it was a disaster. Not to say it wouldn't work if done some other way. If it mattered, I could reboot in hopes I get Eth0 back.
I went into the NAS and enabled Round Robin, not knowing I needed to do more.
You also needed to configure the switch to create a static LAG on the two switch ports.
You really do need to know what you are doing with NIC bonding. Fortunately the bonds are will gracefully fail over if you simply remove one of the two ethernet cables.
If it mattered, I could reboot in hopes I get Eth0 back.
It sounds like that is required. Perhaps power down, remove the power cable for a bit, then reattach it and start up. That sometimes helps when clearing BIOS state is needed.
I haven't made a comment here in awhile. The NAS is still unreliable. It might stay available for hours or days and then it drops off the network. I was never able to get a link open to the server to see if I can connect when the server is otherwise unavailable.
Can't fathom what can be happening if a total rebuild didn't fix the problem and no sign of a hardware error.
