Login or Register to Ask a Question and Join Our Community


Solaris

Strange Multipathing problem

Login to Discuss or Reply to this Discussion in Our Community

 
Thread Tools Search this Thread
Operating Systems Solaris Strange Multipathing problem
# 1  
Old 09-13-2005
Strange Multipathing problem
We have loads of Sparc solaris 8 boxes. Each one has 2 physical interfaces and 2 virtuals set up for multipathing (See ifconfig below). Problem is that I cannot ping the qfe0 interface IP from the local machine. I can ping all other local interfaces, but not 10.161.19.5. I can ping it from other machines, just not the local server. What is also strange is that I am able to ping other machines with the exact same configuration.

Has anyone seen this problem or can anyone explain what exactly happens when you ping a local interface? Does the packet ever reach layer 2, or does it just remain in the kernel since it's a local IP?


server> ifconfig -a
lo0: flags=1000849<UP,LOOPBACK,RUNNING,MULTICAST,IPv4> mtu 8232 index 1
inet 127.0.0.1 netmask ff000000
eri0: flags=9040843<UP,BROADCAST,RUNNING,MULTICAST,DEPRECATED,IPv4,NOFAILOVER> mtu 1500 index 2
inet 10.161.19.4 netmask ffffffe0 broadcast 10.161.19.31
groupname tpfo
eri0:1: flags=1000843<UP,BROADCAST,RUNNING,MULTICAST,IPv4> mtu 1500 index 2
inet 10.161.19.6 netmask ffffffe0 broadcast 10.161.19.31
qfe0: flags=9040843<UP,BROADCAST,RUNNING,MULTICAST,DEPRECATED,IPv4,NOFAILOVER> mtu 1500 index 3
inet 10.161.19.5 netmask ffffffe0 broadcast 10.161.19.31
groupname tpfo
qfe0:1: flags=1000842<BROADCAST,RUNNING,MULTICAST,IPv4> mtu 1500 index 3
inet 10.161.19.7 netmask ffffffe0 broadcast 10.161.19.31
server> ping 10.161.19.5
no answer from 10.161.19.5
server>
# 2  
Old 09-13-2005
i would say that's because you set this interface to deprecated. Addresses marked as deprecated will not be used as source address for outbound packets unless either there are no other addresses available on this interface or the application has bound to this address explicitly.
but it works on my config:
Code:
root@jumpy # ifconfig -a
lo0: flags=1000849<UP,LOOPBACK,RUNNING,MULTICAST,IPv4> mtu 8232 index 1
        inet 127.0.0.1 netmask ff000000
dmfe0: flags=9040843<UP,BROADCAST,RUNNING,MULTICAST,DEPRECATED,IPv4,NOFAILOVER> mtu 1500 index 2
        inet 192.168.133.102 netmask ffffff00 broadcast 192.168.133.255
        groupname PRESSY
        ether 0:3:ba:14:e1:c9
dmfe0:1: flags=1000843<UP,BROADCAST,RUNNING,MULTICAST,IPv4> mtu 1500 index 2
        inet 192.168.133.100 netmask ffffff00 broadcast 192.168.133.255
dmfe1: flags=9040843<UP,BROADCAST,RUNNING,MULTICAST,DEPRECATED,IPv4,NOFAILOVER> mtu 1500 index 3
        inet 192.168.133.103 netmask ffffff00 broadcast 192.168.133.255
        groupname PRESSY
        ether 0:3:ba:14:e1:ca
dmfe1:1: flags=1000843<UP,BROADCAST,RUNNING,MULTICAST,IPv4> mtu 1500 index 3
        inet 192.168.133.101 netmask ffffff00 broadcast 192.168.133.255
root@jumpy # ping 192.168.133.102
192.168.133.102 is alive
root@jumpy # ping 192.168.133.103
192.168.133.103 is alive
root@jumpy # route get 192.168.133.102
   route to: jumpy2
destination: jumpy2
       mask: 255.255.255.255
  interface: dmfe0
      flags: <UP,DONE>
 recvpipe  sendpipe  ssthresh    rtt,ms rttvar,ms  hopcount      mtu     expire
       0         0         0         0         0         0      8232         0
root@jumpy # route get 192.168.133.103
   route to: jumpy3
destination: jumpy3
       mask: 255.255.255.255
  interface: dmfe1
      flags: <UP,DONE>
 recvpipe  sendpipe  ssthresh    rtt,ms rttvar,ms  hopcount      mtu     expire
       0         0         0         0         0         0      8232         0
root@jumpy #

take a look at this, perhabs it helps:
https://www.unix.com/solaris/20665-interfaces-do-not-failover.html

could you post you ipmp settings... and a "netstat -r" output
# 3  
Old 09-13-2005
I also have another machine that is able to ping its own interface. Both were jumpstarted the same way. Do you know if the ping packets reach layer 2? I'm assuming not, because snoop doesn't see anything. Just want to know if it's worth it to turn on packet snooping on the switch.

Code:
# netstat -rn

Routing Table: IPv4
  Destination           Gateway           Flags  Ref   Use   Interface
-------------------- -------------------- ----- ----- ------ ---------
5.128.40.32          5.128.40.45           U        1   1796  qfe1
10.161.19.0          10.161.19.6           U        1   3511  eri0:1
10.161.19.0          10.161.19.6           U        2      0  eri0
10.161.19.0          10.161.19.6           U        1    878  qfe0
5.0.0.0              5.128.40.33           UG       1    322
224.0.0.0            10.161.19.6           U        1      0  eri0:1
default              10.161.19.3           UG       1   2783
127.0.0.1            127.0.0.1             UH      2934882107  lo0
# cat /etc/default/mpathd
#
#ident  "@(#)mpathd.dfl 1.1     00/01/03 SMI"
#
# Time taken by mpathd to detect a NIC failure in ms. The minimum time
# that can be specified is 100 ms.
#
FAILURE_DETECTION_TIME=1000
#
# Failback is enabled by default. To disable failback turn off this option
#
FAILBACK=yes
#
# By default only interfaces configured as part of multipathing groups
# are tracked. Turn off this option to track all network interfaces
# on the system
#
TRACK_INTERFACES_ONLY_WITH_GROUPS=yes

Last edited by tjlst15; 09-13-2005 at 03:45 PM..
# 4  
Old 09-13-2005
hmm, which layer are you talking about? a ping is an icmp protocol (type 8, echo request) and in the iso/osi model on the 4th layer, the transport layer.. but anyway my routing table looks a bit different, you are missing one "dummy" interfaces but you have the qfe1 active which isn't configured?
mine:
Code:
root@jumpy # netstat -r

Routing Table: IPv4
  Destination           Gateway           Flags  Ref   Use   Interface
-------------------- -------------------- ----- ----- ------ ---------
192.168.133.0        jumpy                U         1     67  dmfe0:1
192.168.133.0        jumpy1               U         1     56  dmfe1:1
192.168.133.0        jumpy1               U         1      0  dmfe0
192.168.133.0        jumpy1               U         1      0  dmfe1
BASE-ADDRESS.MCAST.NET jumpy                U         1      0  dmfe0:1
default              192.168.133.50       UG        1     42
localhost            localhost            UH        2     15  lo0
root@jumpy #

with the ipmp configuration i was meant the /etc/hostname.*
have you ever tried to switch the active adapter and switch back, just for testing purpose ( see: man if_mpadm ) also try the "route get 10.161.19.5" command to see which interface solaris wants to use
gP
Login or Register to Ask a Question

Previous Thread | Next Thread

10 More Discussions You Might Find Interesting

1. Solaris

MPxIO Multipathing Problem

Hi, I believe this is an OS multipathing issue - Solaris 10 We have x2 Dual port 8GB Qlogic HBA's installed in our SOlaris 10 host. HBA0 (IOU0) Port WWN's - 09f4 (pci@3,700000/SUNW,qlc@0) 09f5 (pci@3,700000/SUNW,qlc@0,1) HBA1 (IOU1) Port WWN's... (2 Replies)
Discussion started by: jamba1
2 Replies

2. Solaris

Multipathing - problem

Hello, I turned on the server multipathing: # uname -a SunOS caiman 5.10 Generic_141444-09 sun4v sparc SUNW,T5140 stmsboot -D fp -e And after a reboot the server, multipathing is not enable: # stmsboot -L stmsboot: MPxIO is not enabled stmsboot: MPxIO disabled # ls /dev/dsk... (4 Replies)
Discussion started by: bieszczaders
4 Replies

3. Shell Programming and Scripting

Strange Problem

Hi All, I am encountering an unexpected problem while running my process, using daemon placed in /etc/init.d/. I have a process that is opening shared library using dlopen. When I run my process on the command line, it is able to open the library. But when I used daemon to run my process, I... (0 Replies)
Discussion started by: qwerty-1
0 Replies

4. Solaris

Veritas Multipathing problem.

Hi, Basically the original configuration on my Solaris 9 server was two LUNs configured as Veritas file systems that were connected to a NetApp filer (filer1). These two LUNs are still configured on the server - but are not being used. They are there as a backup just in case the new... (0 Replies)
Discussion started by: sparcman
0 Replies

5. Shell Programming and Scripting

Strange problem.

Well, my script started off to do what i wanted. Now, i think its not recognizing the pattern so its not moving anything. What i have to do is execute my script command for the move to take effect. So i did that and yayy it worked. Strange thing is that my DESTDIR was empty to begin with.... (2 Replies)
Discussion started by: oxoxo
2 Replies

6. Shell Programming and Scripting

Strange problem

I am using SunOS 5.9 and I don't know why all my commands are getting executed as if an extra 'enter' has been pressed. What could be the reason and how to correct it? Please help. Asty (2 Replies)
Discussion started by: Asty
2 Replies

7. Linux

very strange problem

I have installed Fedora Core on a Toshiba Satellite Pro4600 laptop recently I have experienced a rather mysterious problem if I touch anything specially the keyboard or mouse I see this stuff “67yujhnmyyy” straight away some time it won't stop for while like this... (5 Replies)
Discussion started by: kemobyte
5 Replies

8. UNIX for Dummies Questions & Answers

strange...problem

Hi.. Some of my application were not running properly due to lack of virtual memory.....so wht i did add one free harddisk as swap file system...and increased the swap memory.. But since than my root file system is showing 100% full thr is no space left...is thr any link between these two..... (1 Reply)
Discussion started by: Prafulla
1 Replies

9. UNIX for Dummies Questions & Answers

Strange Problem

I did an installation of Redhat 7.1 and it seem to install smoothly. I go and reboot and I get the following message: "INIT: Id "x" respawing too fast: disabled for 5 minutes" The screen flashes 4 times every 10 seconds or so. I can't seem to figure why this does this. It might be a simple... (2 Replies)
Discussion started by: forjohnny
2 Replies

10. UNIX for Dummies Questions & Answers

Strange problem.Please Help !

I¡¯m a network operator, mine is an IBM PC server 320, operating system SCO unix 3.2v4.2, triton 3.1 of Baan. Recently, my server went dead every a few hours, no sign & signal shows malfunction suddenly. It looks like a sudden power failure, but the indicator of main power supply is on. Normal... (1 Reply)
Discussion started by: lyhsm
1 Replies
Login or Register to Ask a Question