Strange Connection Probs Bad Eth0 ?

[stevenriz]

Hello, we are trying to diagnose a fairly new issue with a Sol10 machine. It seems to be sporadic in a way. An application on a Fedora machine has been using this particular Solaris machine and all of a sudden it is getting connection timeouts. These machines are on the same subnet. Pings are consistant to a certain few machines but can only ping another machine once in a while. I don't understand. I thought it was an eth0 card on the Solaris machine but it consistently pings some machines, it is just one particular one that it cannot ping consistently. You'll see below this machine is having trouble with .193 and .206

-bash-3.00# ping 192.168.1.206
no answer from 192.168.1.206
-bash-3.00# ping 192.168.1.200
192.168.1.200 is alive
-bash-3.00# ping 192.168.1.200
192.168.1.201 is alive
-bash-3.00# ping 192.168.1.202
192.168.1.202 is alive
-bash-3.00# ping 192.168.1.202
192.168.1.203 is alive
-bash-3.00# ping 192.168.1.200
192.168.1.205 is alive
-bash-3.00# ping 192.168.1.200
192.168.1.206 is alive
-bash-3.00# ping 192.168.1.193
no answer from 192.168.1.193
-bash-3.00#

 

[Annihilannic]

That's very odd, how come when you ping some addresses it frequently replies that another address is alive?

What happens if you ping -s 192.168.1.193, do they randomly succeed and/or fail.

What is your subnet mask?

Annihilannic.

 

[stevenriz]

here is ifconfig and below is the ping -s. Now let me narrate here. The first ping sat there for a couple minutes before it started pinging. Then it seemed fine. These systems are on the same subnet. Even on the same switch! other thoughts?

-bash-3.00# ifconfig -a
lo0: flags=2001000849<UP,LOOPBACK,RUNNING,MULTICAST,IPv4,VIRTUAL> mtu 8232 index 1
inet 127.0.0.1 netmask ff000000
hme0: flags=1000843<UP,BROADCAST,RUNNING,MULTICAST,IPv4> mtu 1500 index 2
inet 192.168.1.155 netmask ffffff00 broadcast 192.168.1.255
ether 8:0:20:c6:1a:57
-bash-3.00#
-bash-3.00# ping -s 192.168.1.193
PING 192.168.1.193: 56 data bytes
64 bytes from 192.168.1.193: icmp_seq=0. time=0.445 ms
64 bytes from 192.168.1.193: icmp_seq=1. time=5.02e+03 ms
64 bytes from 192.168.1.193: icmp_seq=2. time=4.02e+03 ms
64 bytes from 192.168.1.193: icmp_seq=3. time=3.02e+03 ms
64 bytes from 192.168.1.193: icmp_seq=4. time=2.02e+03 ms
64 bytes from 192.168.1.193: icmp_seq=5. time=1.02e+03 ms
64 bytes from 192.168.1.193: icmp_seq=6. time=25.1 ms
64 bytes from 192.168.1.193: icmp_seq=7. time=0.399 ms
^C
----192.168.1.193 PING Statistics----
9 packets transmitted, 9 packets received, 0% packet loss
round-trip (ms) min/avg/max/stddev = 0.307/1.68e+03/5.02e+03/1.9e+03

 

[stevenriz]

This one sat for about 15 minutes and I finally cancelled it. This being just one minute after the previous message I left.

-bash-3.00# ping -s 192.168.1.193
PING 192.168.1.193: 56 data bytes
^C
----192.168.1.193 PING Statistics----
626 packets transmitted, 0 packets received, 100% packet loss
-bash-3.00#

 

[Annihilannic]

Those are some seriously slow pings... do your switch port speed settings (speed and duplex) match those on your systems?

If the problems are specific to the Solaris system (i.e. all the other systems can ping each other fine) I suspect you may need to set the hme interface to full duplex using the ndd command and/or the appropriate entries in /etc/system. Try the former first, and if that works make the change permanent by doing the latter.

If that's not the problem there's something wrong with your cables, switch or... something.

Annihilannic.

 

[stevenriz]

Thanks Annihilannic.... We have all unmanaged switches. They appear to be a 100full but I can't be positive. I changed ports with no difference in performance. I will change cables since that is a pretty easy thing to do. This particular system has had problems recently with bad performance. At that time the system was running Sol8. I since upgraded it thinking it will fix the problem. Apparently it did not. Maybe cables. Maybe bad eth port. I will keep trying things and let you know what I find. Thanks for the suggestions.!!
Steve