Linux server packet loss - stopped occassionally

[Net Llama!]

On 08/24/03 00:34, Swapana Ghosh wrote:

> Hi 
> 
>          Few days we are having problem with one of
> our clients's server(linux 7.1)...
> Within last few days, we had to reboot the server
> atleast 12 times...
> 
>          When the server stopped to work - we are
> checking the *traceroute*..
> We are getting that after prceeding upto 10/12 hubs
> the server is not sending the package... We contacted
> their server maintainance engineers , they are telling
> that they have no problem with the router
> /network...They are maintaining  so many
> servers - if they have problem with router/network
> then others' servers should have been also
> under same condition!!!
> 
>          Now their engieers without rebooting the
> server just did as follows:
> 
> #ifconfig eth0 down
> #ifconfig eth0 up
> #route add default gw 216.240.145.1
> 
>          Then the server started working....
> 
> Here is the servers ifcfg-eth0 file cofiguration
> 
> DEVICE=eth0
> ONBOOT=yes
> BOOTPROTO=static
> IPADDR=216.240.145.47
> NETMASK=255.255.255.0
> GATEWAY=216.240.145.1
> 
> Can anybody help me out to diagnose the problem?

Have you checked any logs to see if there are errors?  Are there any 
ipchain/ipfilter rules in place?  Can you ssh to the box during the times 
that it can't see the outside world?  How confident are you that the box 
hasn't been compromised?

-- 
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
L. Friedman                       	       netllama at linux-sxs.org
Linux Step-by-step & TyGeMo: 		    http://netllama.ipfox.com

   6:20am  up 8 days, 18:47,  1 user,  load average: 0.00, 0.00, 0.00