Hi Jochen,

Jochen Friedrich wrote:

>>On Dell em64t systems running the amd64 port of Debian, interfaces statistics
>>polled via snmp from remote hosts stop incrementing.  I haven't tested
>>on other hardware so can only speak to Dell hardware at the moment.  But
>>the bug shows up on every system running Debian's amd64 port.  Systems
>>running the i386 port with 64bit kernels DO NOT exhibit this problem.
> 
> I'm running snmpd on Alpha, which is also 64bit, and i can't reproduce
> this. Could you
> run "cat /proc/net/dev" on your machine after snmp stops working?

On a system running 5.1.2:

$ cat /proc/net/dev
Inter-|   Receive                                                |  Transmit
 face |bytes    packets errs drop fifo frame compressed multicast|bytes
   packets errs drop fifo colls carrier compressed
    lo:16614789  400707    0    0    0     0          0         0
16614789  400707    0    0    0     0       0          0
  eth0:1815249735 9609173    0    0    0     0          0         0
7193137948 6360542    0    0    0     0       0          0
  eth1:335565863 3770061    0    0    0     0          0         0
10131073060 7034022    0    0    0     0       0          0
  sit0:       0       0    0    0    0     0          0         0
 0  0    0    0    0     0       0          0

That same system when polled remotely shows:

$ snmpwalk -v 1 -c community hostname ifInOctets
IF-MIB::ifInOctets.1 = Counter32: 16614789
IF-MIB::ifInOctets.2 = Counter32: 1815250769
IF-MIB::ifInOctets.3 = Counter32: 335565863
IF-MIB::ifInOctets.4 = Counter32: 0

$ snmpwalk -v 1 -c community hostname ifOutOctets
IF-MIB::ifOutOctets.1 = Counter32: 16614789
IF-MIB::ifOutOctets.2 = Counter32: 16614789
IF-MIB::ifOutOctets.3 = Counter32: 16614789
IF-MIB::ifOutOctets.4 = Counter32: 0

The /proc counters appear to be incrementing properly, as do the
ifInOctets.  But the ifOutOctet counters are stuck at the values above.
 I can send you the same information for 5.2.1 as soon as that system
starts exhibiting the problem (which I believe is only a matter of
time).  Also, I haven't tried polling the data locally as I do not have
snmpget/snmpwalk installed on these machines.  I'll try that later
today.  The polling system used above is polling a lot of other machines
(it's sarge i386 on x86) so I'm pretty sure it's not at fault.

Thanks.
Regards,
Ed


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Reply via email to