It shouldn't be either/or; it should be both. That is why I suggested ethernet-level diagnostics. Have you looked at dmesg? You may find error messages reported by the ethernet driver there.

--- rod.

Yes, those CIFS errors came from dmesg.

I had a few of these in dmesg, but the last line is where we're at right now:

(mods: I reported this to be moved a few days ago... can someone please move it back to its original forum?)

Please accept my apology for the delay. We are experiencing a heavier than usual call volume. Your re-move will be made as soon as one becomes available. Thank you for your patience.

Woops, wrong subforum.

Thanks for moving...

Can someone with perhaps some server knowledge let me know why my CIFS mounts are failing? I'm not sure if it's an Ethernet issue, and I'm open to figuring this out, since as you may see in my message -- code is failing. It's not a code issue though.

The code is not working, but you can solve your problem by fixing the network. Start by dis-connecting and re-connecting, one at a time and retesting for errors, all cables that touch the path between the hosts involved. If the ethernet interface is removable, remove it, and re-insert it (power down while replacing cards). Power cycle all elements that touch the path between the devices involved. If possible, re-arrange components like switches and hubs, and see if the fault follows one of them. Try different, known-good components of all sorts, if possible. Look for cables that may have been damaged by heat, physical damage such as rodent bites, kinks and knots, scrapes and abrasion. If cables have been coiled up into neat rolls, undo those as it may be nicer looking, but can hamper performance. Try to reduce the scope of the network to only the hosts that are misbehaving. It may be that the fault exists with another device on the network, and neither of the hosts that exhibit the symptoms. Be systematic; make one or two changes at a time, only. If the problem goes away, you want to know which thing was involved.
The problem does not seem to be a CIFS issue, as the network interface itself is going down and up. Is it possible that the driver is the wrong one? There are a few Intel e1000-like drivers. Its possible that the installed driver supports a subset of what is required to make the hardware function correctly.

--- rod.

@theNbomr - this isn't a complex setup. There's one cable from both machines and they each go directly into a relatively new hub. These are personal boxes, not a complex server setup.

The code has been working and is "final" since 2004. I know it's "out of date" by our standards but I'm pretty sure that it's not a code issue. To be even clearer since I apparently missed this point, this problem only began when I upgraded my CIFS machine from Windows XP to Windows 7, since I never seemed to have experienced this when running on Windows XP, at least not to my awareness. No network cables were harmed in the upgrade of my machine. It's a Windows 7 issue/mount problem if anything, not a coding issue. Or so I'm convinced -- because all signs point to that. Shrug. Maybe I'm wrong, but I don't really know.

According to dmesg, the NIC is currently stable and nothing has been altered. Still, it's not working. Sometimes it does, and sometimes it doesn't, but lately it doesn't.

As far as the driver, I'm open to trying, but this was an auto-detection and installation, so should I assume Fedora selected the wrong driver? Again, the driver and Linux box haven't been touched; the only thing that was changed was an upgrade to Windows 7.

I guess this is where nobody has the answers

You seem to have the answer already.

--- rod.

And I'm asking for a solution. Tips?

Okay, trying again.

Does anyone know why a Windows 7 mount via CIFS on a Linux (Fedora 14) box would fail being acknowledged as a valid mount? There are no problems reading, writing, or executing files. The problem is the way the OS sees the mount. Information is in the messages above and the pages preceding, since I haven't really gotten anywhere....

....go!

Maybe this would be something already considered, but is it possible the answer (or more precisely, the problem) lies in the above snippet?

I don't use Windows and have never used 7 at all, ever, and haven't ever used CIFS (or Fedora for that matter!) so I don't follow how well MS stuff is working generally. But, would it be possible to test this situation with an XP machine in place of the Win7 machine, and see if the problem goes away?

Yeah, I guess I could. I guess I'm not fully clear on the problem, and was wondering if it was just me using Windows 7 or if it's a little more complicated.

Questions could be:
* Is it my specific Windows 7 installation?
* Is there something wrong with Windows 7 in general?
* And the more obvious contender that you suggested: is it just generically XP vs. Win7? (which is related to the aforementioned questions)
* etc.

That's kind of where I was headed -- I appreciate that, and that's easier to test. I'll probably try a few things to see if I get anything from this.

TY!

Okay, hmm, oddly it's not working on the XP machine either or a second Windows 7 machine that I tested. Maybe I just missed this all around...

So I'm still at the point where the OS is not seeing these valid directories as directories, and yet different results are being seen when running code that has OS calls. It will show the proper "d" (directory) designation when running ls -la, though.

Are there any other things I should try? Right now, the network seems to be fine, so I'm not sure that's a problem I should explore, and there were no CIFS errors in dmesg when I ran the tests.

hmmmmm