C6500 occasionally suffers from a large number of packet losses！

[Rps-Cheers]

HI,

We have a C6509 switch, which has been running for more than 13 years. The version it runs is Version 12.2(18)SXF6.
Recently, we encountered a problem, and sometimes there will be a lot of packet loss on this station. And the device will appear stuck when we log in to the device (probably due to packet loss). When packet loss is severe, there is basically a 50% packet loss rate.

specific situation:
-Maximum traffic on the link is 50-70Mbps
-The interface does not see packet loss or error packet count increase
-No high memory
-No high CPU
-The Module status on the device is also normal
-No abnormal logging
-We tried to replace the interface, cable, and line card. The fault disappeared for a few days, but it appeared again.
-Sometimes, we shut down the interface, and it can work normally when it is opened the next day, but similar phenomena will happen again.

Is the device running for too long, can the restart be resolved?
Is the reason for the older software version causing this problem?
Do you have any suggestions, how can I troubleshoot this problem?

Thx

[Leo Laohoo]

SXJ train is the most well-known because it is very, very stable.

[Rps-Cheers]

Hi Leo

thanks for your response.
As you said, maybe this is a software problem?

[Rps-Cheers]

Hi Leo

I didn't find the SXJ version? Only SXH and SXF can be selected on the Cisco IOS download page

[Leo Laohoo]

---

@Rps-Cheers wrote:
maybe this is a software problem?

---

If the chassis has an uptime of >13 years, then I'd recommend it is time to get the firmware upgraded.

[Rps-Cheers]

I think this is a good suggestion. I don’t know what the end user thinks
Thanks & BR

[Rps-Cheers]

In addition to the inspections I have done above, are there any other troubleshooting suggestions?
Thx

[Joseph W. Doherty]

It might also be helpful if you noted the all the hardware particulars beyond its a 6509.

BTW, take note of Leo's comment, SXF is rater dated and w/o looking I'm unsure how (18)F6 stands in its software train either.

[Rps-Cheers]

How to understand "noted the all the hardware particulars beyond its a 6509"
Other devices are not abnormal. When disconnecting the 6509, everything is normal
Is SXJ very stable as Leo said, should I choose an SXJ version to upgrade?

Thx

[Joseph W. Doherty]

How to understand "noted the all the hardware particulars beyond its a 6509"

i.e.
What model 6509, -E or not.
Sup?
Line cards?
Port speed with issue?

[Rps-Cheers]

Not -E, and Sup is WS-SUP720-3B,LC is WS-X6724-SFP, 1000Mbps(1G)

[Giuseppe Larosa]

Hello @Rps-Cheers ,

to see your hardware's list post

show module

in this way it is possible to see the type of supervisor and the type of linecards and if there are daughter cards like DFC installed.

Can you post also

show version | inc uptime

Hope to help

Giuseppe

[Rps-Cheers]

HI Giuseppe

Please check the attachment information

Thanks a lot

[Giuseppe Larosa]

Hello @Rps-Cheers ,

there are no Tengiga linecards and this is likely the reason why you could stay with this IOS release.

but be aware that the two supervisors have different IOS versions:

7 0012.xxxx.xxxx to 0011.xxxx.xxxz 5.2 8.4(2) 12.2(18)SXF6 Ok
8 0013.xxxx.xxxx to 0011.xxxx.xxxb 5.3 8.4(2) 12.2(18)SXF7 Ok

The end result of this IOS mismatch is that instead of being able to run SSO stateful switchover you have RPR as redudancy strategy in fact the standby supervisor is in state Cold and not HOT standby.

7 2 Supervisor Engine 720 (Active) WS-SUP720-3B SALXXXXXXXX
8 2 Supervisor Engine 720 (Cold) WS-SUP720-3B SALXXXXXXXX

This would mean that in case of failure of primary supervisor the Cold standby will take more time to recover.

The uptime is really impressive more then 13 years.

Whatever IOS version you decide to upgrade to ensure the same version is present on both supervisors.

Hope to help

Giuseppe

[Rps-Cheers]

HI Giuseppe,
This is indeed the case, I have not noticed that the versions of the Active and backup engines are different, but this should not be the root cause of the problem.
Regarding this abnormal situation of packet loss, do you also think it should be upgraded?
If the end user upgrades, the two engines must be upgraded to the same version.

Thank you