Topic: local network crashing

[thecat]

Nov 25 12:16:05 proliant7000 dhcpd: svc: warning: unable to control /service/dhcpd: supervise not running
Nov 25 12:16:06 proliant7000 smb: svc: warning: unable to control /service/nmbd: supervise not running
Nov 25 12:16:06 proliant7000 smb: svc: warning: unable to control /service/smbd: supervise not running

I have no idea why but my local network is constantly going down. My external interface is working. I have tried restarting network - dhcp - dhcpcd - smb etc etc but nothing unables the clients to pick up the local network. The only thing that does is a complete server restart. I noticed these entries in the log file. Could this be where the problem is? If so anyone have any information which might be helpful?

[thecat]

here is some more details hopefully someone might be able to send me in the right direction.

from my logwatch
   ICMP Echo Reply for 192.168.70.66 arrived late or is spurious.: 2 Time(s)
   ICMP Echo Reply for 192.168.70.81 arrived late or is spurious.: 2 Time(s)
   Starting dhcpd succeeded: 24 Time(s)
   Starting dhcpd:: 10 Time(s)
   Stopping dhcpd succeeded: 16 Time(s)
   Stopping dhcpd:: 2 Time(s)
   ^[[60G: 12 Time(s)
   send_packet: Operation not permitted: 20 Time(s)
   svc: warning: unable to control /service/dhcpd: supervise not running: 2 Time(s)
   unexpected ICMP Echo Reply from 150.101.180.192: 8 Time(s)

DHCP Server Listening On:
   Socket/eth0/192.168.70.0: 2 Time(s)
   Socket/eth1/192.168.70.0: 28 Time(s)

the server is a compaq proliant 7000 with 2 TLAN nics. I have tried running the nics in normal and swapped mode, but it makes no difference. I know there is a bug report about the external and internal nics getting confused when but set up with dhcp, but I am not sure if that my problem or not?

[MSmith]

1)  Which SME version?
2)  What happens if you disable one or both of the onboard NICs and put in a cheapo PCI NIC?

[thecat]

it is 6.5 but is heavily modified.

Haven't tried another card, because if I switch the clients to fixed ip's and turn off dhcp all is well. This would seem to indicate to me that it is not an hardware issue.

Thanks for you reply

[thecat]

well it just failed again even with dhcp turned off. Very annoying after going around to 50 pc's and setting up manual ips!

I have tried new network cards and it still crashes.

The day has come but I am bidding sme good bye

[alejandro]

Is there any (misconfigured) pc trying to serve as dhcp too?

[thecat]

thanks for the reply.The 'other' dhcp server idea was what I thought at first but could not find anything. Unless someone has access to my wirless network and is running a dhcp server (very very remote chance - but I am watching).

So far I have tried a total of 4 TLAN cards with a mixture of other TLAN and non TLAN's and all have had the problem. However, last night I replaced all of the TLAN driver based cards with different cards and all seems well, for at least 12 hours.

It really does seem that with the TLAN driver based cards, when both WAN and local LAN are setup using DHCP. The LAN card will for some strange reason, at any time, get confused and instead of giving out IP's will wait to recieve an IP

[alejandro]

did you tried changing dhcp lease time? (shorter) just to see if the problem persist

maybe it's obvious, but.... here it goes: are the APs  configured to  provide neither dhcp service nor mac/ip filtering?
could be usefull knowing a little more of your topology/scenario

[thecat]

no i didn't try changing the lease time, simply because I thought restarting the network /etc/rc.d/init/network restart and or dhcp restart would actually renew the leases.

I have another proliant server with the same cards waiting to put into service. So I might see if I can reproduce the fault tonight.

The AP's are simple repeaters, therefore do not even have dhcp.

I apprecriate your efforts in helping me

[MSmith]

I think the "heavily modified" part might be the cause of your problem ... what happens if you use a stock 6.5 or 7.0 install?

[thecat]

I will try to reproduce the fault on a stock 6.5 machine tonight. I did notice a bug report somewhere (dam if I can find it now) regarding this same problem.

[thecat]

I found something similar with this bug report

http://no.longer.valid/mantis/bug_view_page.php?bug_id=0000071

[thecat]

ok, I installed 6.5 onto another proliant 7000 last night with 2 TLAN driver based network cards. And the results are the same, after the period of time the local LAN card seems to be getting confused and starts to behave as though it is the WAN card.

I'm not certain if heavy transfering of files, is the catalyst or not, because I can crash the network more often than not by doing a server backup. However, I think on several occassions it just happen without any activity (not sure on this - because some of the local pc's could have been doing windows updates etc etc)

I am not sure why this http://no.longer.valid/mantis/bug_view_page.php?bug_id=0000071 was closed in the first place?

[CharlieBrady]

I am not sure why this http://no.longer.valid/mantis/bug_view_page.php?bug_id=0000071 was closed in the first place?

It was closed because someone was asked for more information but they didn't provide it.

[thecat]

would you like anymore information on it? or is reporting bugs on 6.5 now a waste of time?