[Techtalk] troubleshoot slow ping and packets lost

[Joana Botto]

Hi,

I checked the logs, I've copied what I think is relevant below and it
seems I had a NIC fail and it successfully failed over to ce0.
ce1 doesn't appear in cfgadm. I haven't read yet IPMP Sun document
which seems very complete and very big but isn't bonding to prevent
these situations of packets lost? Or maybe there is another obscure
reason behind this slowness and random packets lost? Thank you.

root at tndub001# cat /var/adm/messages
Jun  3 10:55:53 tndev01 in.mpathd[132]: [ID 594170 daemon.error] NIC
failure detected on ce1 of group prod
Jun  3 10:55:53 tndev01 in.mpathd[132]: [ID 832587 daemon.error]
Successfully failed over from NIC ce1 to NIC ce0
Jun  3 10:56:07 tndev01 in.mpathd[132]: [ID 620804 daemon.error]
Successfully failed back to NIC ce1
Jun  3 10:56:07 tndev01 in.mpathd[132]: [ID 299542 daemon.error] NIC
repair detected on ce1 of group prod
Jun  9 13:36:39 tndev01 sshd[20202]: [ID 800047 auth.crit] fatal: Read
from socket failed: Connection reset by peer

root at tndub001# more /etc/hostname.ce0
tndev01a netmask + group prod -failover  up  addif tndev01 netmask + up
root at tndub001# more /etc/hostname.ce1
tndev01b netmask +  group prod -failover  up

# Internet host table
#
127.0.0.1       localhost
10.160.30.56    tndev01   tndev01        tndev001 tndub001       loghost
10.160.30.55    tndev01a
10.160.30.54    tndev01b

root at tndub001# cfgadm
Ap_Id                          Type         Receptacle   Occupant     Condition
c0                             scsi-bus     connected    configured   unknown
usb0/1                         unknown      empty        unconfigured ok


Regards,
Joana Botto


On Tue, Jun 9, 2009 at 9:29 PM, Joana Botto<joana.botto at gmail.com> wrote:
> Hi Carol,
>
> Thank you for your tips. I've been busy with other IT problems but
> I'll look at this issue this week and be sure to update this email,
> because I'll probably have more questions =)
>
> Regards,
> Joana Botto
>
> On Mon, Jun 8, 2009 at 4:05 PM, Carol Williams<carolswilliams at gmail.com> wrote:
>> Joana,
>> Have you done a tcpdump from both sides to see what the traffic looks like?
>> I see you are using IPMP, are you getting any failure messages in /var/logs
>> about an interface being deprecated?  Also, how does your routing table
>> look?  Do you have test IP's configured on each of the interfaces for probed
>> based failures?
>>
>> Carol Williams
>>
>> --
>> If you can spend a perfectly useless afternoon in a perfectly useless
>> manner, you have learned how to live.  -- Lin Yutang
>>
>