| View previous topic :: View next topic |
| Author |
Message |
gncuster
n00b
Joined: 16 Aug 2002
Posts: 20
|
 Posted: Mon Aug 26, 2002 11:01 pm Post subject: Wierd Network Preformance |
 |
|
Hey all,
Recently my machine has been acting very strangly. I timeout when resolving most dns queries; and I time out for many other simple network tasks. It seems like my computer is holding all outbound packets in a que and waiting to send them. I am running the 2.4.19 gentoo kernel, 1.4_beta, with a sis900 built in network card.
Here is the output from a pinging session. My gentoo machine is 10.0.0.8, the 10.0.0.5 box is a debian server. They are on the same 10/100 switch so the lag should be _very_ small:
| Code: | rd@rd-dev:~$ ping 10.0.0.8
PING 10.0.0.8 (10.0.0.8): 56 data bytes
64 bytes from 10.0.0.8: icmp_seq=0 ttl=64 time=6993.4 ms
64 bytes from 10.0.0.8: icmp_seq=1 ttl=64 time=5997.9 ms
64 bytes from 10.0.0.8: icmp_seq=2 ttl=64 time=4998.6 ms
64 bytes from 10.0.0.8: icmp_seq=3 ttl=64 time=3998.9 ms
64 bytes from 10.0.0.8: icmp_seq=4 ttl=64 time=2998.9 ms
64 bytes from 10.0.0.8: icmp_seq=5 ttl=64 time=1999.0 ms
64 bytes from 10.0.0.8: icmp_seq=6 ttl=64 time=998.9 ms
64 bytes from 10.0.0.8: icmp_seq=19 ttl=64 time=0.3 ms
64 bytes from 10.0.0.8: icmp_seq=7 ttl=64 time=11999.8 ms
64 bytes from 10.0.0.8: icmp_seq=8 ttl=64 time=10999.9 ms
64 bytes from 10.0.0.8: icmp_seq=9 ttl=64 time=10000.0 ms
64 bytes from 10.0.0.8: icmp_seq=10 ttl=64 time=9000.0 ms
64 bytes from 10.0.0.8: icmp_seq=11 ttl=64 time=8000.1 ms
64 bytes from 10.0.0.8: icmp_seq=12 ttl=64 time=7000.2 ms
64 bytes from 10.0.0.8: icmp_seq=13 ttl=64 time=6000.2 ms
64 bytes from 10.0.0.8: icmp_seq=14 ttl=64 time=5000.3 ms
64 bytes from 10.0.0.8: icmp_seq=15 ttl=64 time=4000.3 ms
64 bytes from 10.0.0.8: icmp_seq=16 ttl=64 time=3000.4 ms
64 bytes from 10.0.0.8: icmp_seq=17 ttl=64 time=2000.5 ms
64 bytes from 10.0.0.8: icmp_seq=18 ttl=64 time=1000.5 ms
64 bytes from 10.0.0.8: icmp_seq=34 ttl=64 time=0.3 ms
64 bytes from 10.0.0.8: icmp_seq=20 ttl=64 time=13999.8 ms
64 bytes from 10.0.0.8: icmp_seq=21 ttl=64 time=12999.9 ms
64 bytes from 10.0.0.8: icmp_seq=22 ttl=64 time=11999.9 ms
64 bytes from 10.0.0.8: icmp_seq=23 ttl=64 time=11000.0 ms
64 bytes from 10.0.0.8: icmp_seq=24 ttl=64 time=10000.0 ms
64 bytes from 10.0.0.8: icmp_seq=25 ttl=64 time=9000.1 ms
64 bytes from 10.0.0.8: icmp_seq=26 ttl=64 time=8000.2 ms
64 bytes from 10.0.0.8: icmp_seq=27 ttl=64 time=7000.2 ms
64 bytes from 10.0.0.8: icmp_seq=28 ttl=64 time=6000.3 ms
64 bytes from 10.0.0.8: icmp_seq=29 ttl=64 time=5000.3 ms
64 bytes from 10.0.0.8: icmp_seq=30 ttl=64 time=4000.4 ms
64 bytes from 10.0.0.8: icmp_seq=31 ttl=64 time=3000.5 ms
64 bytes from 10.0.0.8: icmp_seq=32 ttl=64 time=2000.5 ms
64 bytes from 10.0.0.8: icmp_seq=33 ttl=64 time=1000.6 ms
--- 10.0.0.8 ping statistics ---
38 packets transmitted, 35 packets received, 7% packet loss
round-trip min/avg/max = 0.3/6028.3/13999.8 ms |
Its worth noting that all of the packets in a count-down group arived at the same time. Does anyone know what the cause of this might be?
--Nate |
|
| Back to top |
|
 |
rac
Bodhisattva
Joined: 30 May 2002
Posts: 6553
Location: Japanifornia
|
| Posted: Mon Aug 26, 2002 11:03 pm Post subject: |
|
|
I'm sure everybody is getting tired of my suggesting this, but does the behavior persist even using vanilla kernel sources?
_________________
For every higher wall, there is a taller ladder |
|
| Back to top |
|
|
gncuster
n00b
Joined: 16 Aug 2002
Posts: 20
|
| Posted: Mon Aug 26, 2002 11:16 pm Post subject: re: vanilla kernel |
|
|
Rac,
> I'm sure everybody is getting tired of my suggesting this
Does this mean you have seen this problem before?
I will look into it, part of the problem I am having is that this problem has reduced my http sessions to a dialup like pace 
--Nate |
|
| Back to top |
|
|
rac
Bodhisattva
Joined: 30 May 2002
Posts: 6553
Location: Japanifornia
|
| Posted: Mon Aug 26, 2002 11:27 pm Post subject: Re: re: vanilla kernel |
|
|
| gncuster wrote: | Rac,
> I'm sure everybody is getting tired of my suggesting this
Does this mean you have seen this problem before? |
Not this specific one, but it's sort of become my default response whenever somebody is having unusual problems and using gentoo-sources.
When I first saw your post, I was ready to suggest that it was a reverse-DNS problem, but as I read it, I saw you were having trouble even using raw IP addresses, so it can't be that.
The Gentoo-patched kernel does all sorts of tweaking of the scheduler, and I don't think it has really gotten enough testing on a broad variety of hardware to have really gotten all of the kinks out. Of course, this is all going to change very soon, as -r8 and beyond appear to be based on 2.4.19, so there should be much fewer patches.
The latest problem I remember where switching to a vanilla kernel worked was this one.
_________________
For every higher wall, there is a taller ladder |
|
| Back to top |
|
|
gncuster
n00b
Joined: 16 Aug 2002
Posts: 20
|
| Posted: Tue Aug 27, 2002 6:11 am Post subject: kernel issues |
|
|
Roc,
It turns out rebooting the machine fixed whatever ailed it. I tested dns issues before I posted. I am not sure if I should follow this further or not. The idea of rebooting to fix a problem wreaks of windows; its a cop out. But I am at a loss to explain what was happening.
In any case feel free to mark this thread as closed.
--Nate |
|
| Back to top |
|
|
orkid
Tux's lil' helper
Joined: 07 Jun 2002
Posts: 94
Location: Toronto, Canada
|
| Posted: Thu Nov 21, 2002 4:21 pm Post subject: |
|
|
| Well. I've been having a similar problem with my P120 IPMASQ router. After a couple of hours of operation the pings get up to about 2 to 3 seconds. Most of the time it gives me "Bad Data Byte #0" errors though. Don't know if it's related to your problem, but rebooting helps as well. I'm using the r9 gentoo kernel. Perhaps I'll give vanilla a try. |
|
| Back to top |
|
|
orkid
Tux's lil' helper
Joined: 07 Jun 2002
Posts: 94
Location: Toronto, Canada
|
| Posted: Fri Nov 22, 2002 2:54 am Post subject: |
|
|
So vanilla is up and running, and I STILL have the same problems. I really am starting to wonder what this could be...
-Mike |
|
| Back to top |
|
|
|
Networking & Security
