DebianThis forum is for the discussion of Debian Linux.
Notices
Welcome to LinuxQuestions.org, a friendly and active Linux Community.
You are currently viewing LQ as a guest. By joining our community you will have the ability to post topics, receive our newsletter, use the advanced search, subscribe to threads and access many other special features. Registration is quick, simple and absolutely free. Join our community today!
Note that registered members see fewer ads, and ContentLink is completely disabled once you log in.
If you have any problems with the registration process or your account login, please contact us. If you need to reset your password, click here.
Having a problem logging in? Please visit this page to clear all LQ-related cookies.
Get a virtual cloud desktop with the Linux distro that you want in less than five minutes with Shells! With over 10 pre-installed distros to choose from, the worry-free installation life is here! Whether you are a digital nomad or just looking for flexibility, Shells can put your Linux machine on the device that you want to use.
Exclusive for LQ members, get up to 45% off per month. Click here for more info.
i'm having quite the frustrated time with my debian box.
i'm running debian woody (with some bits from testing)
i'm trying to configure as a web/mail/ftp server.
all packages are in and working fine (ftp seems flakey - but that can wait).
i'm running kernel 2.4.27-1-686-smp
i have two NICs - one internal one external. the internal card seems fine. the external card drops out every now and then. if i reboot the box, it seems ok for a moment. 5 minutes later, the external NIC is not pingable, can't see web site, mail access is down. wait another 5min and it might come back again. i've tried three cards, so i've effectively ruled out hardware. there is a firewall, but that doesn't seem to be the problem - besides i can access it with no probs when its up ok.
its not the provider as there are multiple win2k servers running similar setups on the network (point here is to replace those with linux).
nothing jumps out at me in the logs.
i don't know where else to look.
01:02.0 Ethernet controller: Realtek Semiconductor Co., Ltd. RTL-8139 (rev 10)
Subsystem: Surecom Technology EP-320X-R
Flags: bus master, medium devsel, latency 32, IRQ 22
I/O ports at a000 [size=256]
Memory at f8000000 (32-bit, non-prefetchable) [size=256]
Capabilities: [50] Power Management version 2
01:08.0 Ethernet controller: Intel Corp.: Unknown device 1050 (rev 02)
Subsystem: Giga-byte Technology: Unknown device 3013
Flags: bus master, medium devsel, latency 32, IRQ 20
Memory at f8001000 (32-bit, non-prefetchable) [size=4K]
I/O ports at a400 [size=64]
Capabilities: [dc] Power Management version 2
the external card just doesn't want to stay up. if i restart networking, it will work for a while, and then down again... and then maybe back up... you get the picture...
anyway, any suggestions or pointers of what to check would be appreciated.
If you've tried changing NICs, then it sounds more like a driver/module problem.
Does the manufacturer perhaps provide a Linux driver of their own?
You could try searching for the NIC chipset on google to se if anyone else have had the same problems.
It might be possible that the module has to be loaded with some special variables.
If a module bails out there is normally a note written to /var/log/messages. I've experienced this with a PCMCIA card which sometimes switched from 100Mbit to 10 Mbit, which confused the system so the network went down. After some googling I found out how to force the card to 100 Mbit and after that all was good!
Try booting with Knoppix to see what modules it chooses.
If everything else fail, switch to another brand of NIC. After all, NICs are cheap and your time is not...
I would suggest checking the broadcast address, and your routing table.
When you say you've tried three different nic's are they all the same type?
I'm not familiar with Debian - if I changed hardware type in RedHat I would have to run Kudzu to detect the new hardware, or modify teh /etc/modules.conf file to make sure the correct (or compatable) drivers were loaded.
i have tried cables. i have tried different modules for the NICs also. i have a new card coming in that i haven;t tried before soon.
one card is a netgear one which i have using the natsemi module, the other is a generic one using the realtek chipset and using the 8139too module. i have swapped them over in terms over internal/external... both work fine for internal - problems happen on both cards when on the external network.
i have fiddled with broadcast values but i don't know if they're correct. what range should these be??
the NICs haven't been all the same type, though i have tried a couple of netgear ones.
the intel interface is actually onboard. i have disabled this a couple of times and used 2 plain cards instead - no change when using this configuration.
Knoppix is a good idea... i'll try that.
all logs tell me nothing unfortunately...
i'm still cracking away... i just want to get this sorted before i look too stupid and the honchos here think the linux move was a bad idea... i pushed so hard for it ... a little frustrating when i can;t even get the network to stay on for longer than about 5 min.
Hi,
since you are talking about fidelling with the braodcast range.... aren't you supposed to get the whole configuration over DHCP (e.g. as root: dhclient eth1) - i suppose the provider knows how to set the broadcast range.
I believe that your isp should determine your broadcast address.
As you have 255.255.255.240 as your ip netmask then that suggests you have an address range of 16 addresses. Your broadcast address is usually the highest valued address in your range (well it has been for me in the past). I can't say any more than you need to check this with your isp as they have control over the address ranges in use.
As one of the previous posts says, it should be within the range of your subnet. With the 255.255.255.240 netmask you have 16 IP addresses in your subnet. The very first is your subnet address, the second is frequently used as the defauilt gateway (router), and the last is usually the broadcst address. Some installations use the subnet address (the first in your range) also as broadcast address. The 13 remaining addresses inyour subnet can be used for individual machines.
So, the first three bytes "my.external.ip" (using your example above) should be the same for router, all machines, and broadcast.
i think its resolved. first, you were right... my broadcast was in fact wrong.
i also tried a knoppix cd and found that the modules i had were what knoppix picked so i was happy with that. (thats a good one btw that i had not thought of... thanks).
further, i removed one of the gateways. in specifying a gateway for each NIC which was also different for each NIC, it would almost get confused as to which one to use, perhaps even loop around and then dump. hence i suspect the up/down characteristics. i copied the network details off a windows box on with a similar internal/external setup and figured that was right.
i removed the gateway from the internal card and kept just the outside one. can anyone confirm that i should have only one gateway amongst 2 cards???
it appears that the combination of the braodcast and the removal of the gateway on the internal card has done the trick (well so far so good anyway).
i'm still a little unsure, but glad its working.
thanks so much to everyone for chipping in with this. i couldn't have done it without you. its also another feather for linux's cap as my honchos over here were concerned about immediate linux troubleshooting support before the 'get someone in' stage. the community has jumped in to help very nicely. i plan to swap over the other windows servers here and be done with them for good (i hope). sorry... feeling very warm and fuzzy right now....
Get used to the warm fuzzy feeling - especially when you see some of the uptimes you can get (I have an intranet server that has been running for 12 months without any admin!).
Gateways - yes you can have more than one if you wish. Each gateway you wish to use is the means by which you get traffic to another subnet - i.e. the gateway to another network. I have some servers running apache in a web server farm, they have a route off the 'front end' (the default route) to serve the public requests, a route off the 'back end' so that I can administer them and they also have a route onto another subnet - for traffic to/from another office!
A route is merely a way of specifying that all traffic for "this subnet" should go via "here". The gateway being the device addressed "here".
LinuxQuestions.org is looking for people interested in writing
Editorials, Articles, Reviews, and more. If you'd like to contribute
content, let us know.