[ale] Centos 6 system on HP Pavillion - NIC chip appears to be going south - options?

[neal at mnopltd.com (Neal Rhodes)]

Thank you to both Michael and Jim for useful replies.   Now it starts to
get weird. 

Here is the configuration that has worked for years, but started acting
up: 

        AT&T Uverse router -> LAN port
                Sonicwall TZ190 Firewall Router 
                        LinkSys 10/100 Workgroup Switch
                                Centos 6 HP Pavillion - couldn't see
        Sonicwall
                                Windows 7/10 Dell Optiplex -
        intermittently could/could not see Sonicwall
                                Dell Laser Printer - neither system
        could print. 
                                Other stuff


I lost any connectivity to/from the Centos box on Saturday and rebooted
several times and got it back each time temporarily. 

Attempting to isolate, I have re-jiggered the wiring temporarily:  

        AT&T Uverse router -> LAN port
                Sonicwall TZ190 Firewall Router 
                                Centos 6 HP Pavillion - can see
        Sonicwall
                                Windows 7/10 Dell Optiplex - can see
        Sonicwall and Centos
                                Win 10 Acer notebook - can see Sonicwall
        and Centos
                                Dell Laser Printer - All workstations
        can print to it.
                                LinkSys 10/100 Workgroup Switch      
                                    Other stuff
        

And the Centos box has been connected with no interruptions since.
Everything seems to work.   

My suspicion has shifted to the LinkSys 10/100 Workgroup Switch.    Got
to think about how one tests non-managed switch. 


On Mon, 2015-09-14 at 00:00 -0400, Jim Kinney wrote:

> I agree. Script up a tool that checks for connection to the gateway.
> If it fails, stop network, unload module, load module, start network
> and retest conn to gateway. If that works, it proves absolutely
> nothing except the connection will restart.
> 
> 
> On Sep 13, 2015 5:25 PM, "Michael Trausch" <mike at trausch.us> wrote:
> 
>         Check your available modules and use a chip which can be
>         driven. Realtek chips are relatively safe as they use the same
>         ones for a long long time.
>         
>         A diagnostic for you: does removal and reinsertion of module
>         work? That'd be able to eliminate many aspects (but not all)
>         of software causes. I don't believe it's possible to do a
>         power cycle from software on the PCI-e bus, or if suggest that
>         as an interim solution. If you bounce the chip as soon as
>         malfunction is detected most tcp sessions will survive it.
>         
>         Sent from my iPhone
>         
>         > On Sep 13, 2015, at 5:02 PM, Neal Rhodes <neal at mnopltd.com>
>         wrote:
>         >
>         > Any recommendations for an inexpensive no-drama PCI x1 card
>         that works out of the box with Centos 6?
>         
>         _______________________________________________
>         Ale mailing list
>         Ale at ale.org
>         http://mail.ale.org/mailman/listinfo/ale
>         See JOBS, ANNOUNCE and SCHOOLS lists at
>         http://mail.ale.org/mailman/listinfo
> 
> _______________________________________________
> Ale mailing list
> Ale at ale.org
> http://mail.ale.org/mailman/listinfo/ale
> See JOBS, ANNOUNCE and SCHOOLS lists at
> http://mail.ale.org/mailman/listinfo


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.ale.org/pipermail/ale/attachments/20150914/a9b2c83e/attachment.html>