Unsolved
This post is more than 5 years old
7 Posts
0
18209
May 5th, 2014 08:00
PowerConnect 62xx crash.Error: VoipPktIntercept(): message send failed
Dear support,
we are experiencing this issue for years. Switches powerConnect 6224 and 6248 occasionally suffer lost of performance and finally get stuck. In the log we can read several messages like this:
VOIP[103498576]: voip_control.c(179) 10946 %% VoipPktIntercept(): message send failed
The only way to recover the switches is rebooting from console or hardware reboot.
We have tried to update the firmware, disable auto VoiP, replace the units, etc. without success.
Now the units are two stacked 6224 with 3.3.4.1 firmware and RSTP.
We have similar configurations with the same firmware at other sites without issues. Any help would be appreciated.
Thanks,
No Events found!
itelcom
7 Posts
0
May 6th, 2014 03:00
Thank you, we will upgrade to the latest firmware as soon as possible. The network configuration of the different sites is very similar, with different models of phones and softphones in each site, so probably this is not the cause.
Different items from other sites are ip cams, and a large access stack of 8 x powerconnect 3548 connected to the 6224.
Best regards,
David
itelcom
7 Posts
0
May 6th, 2014 08:00
That's the problem. We don't know when the error is going to happen even how often it happens. Last issue happened a month ago, but the previous one 6 months ago!
We are not able to find any documentation about the error message to try to discover the cause. Could you help us with this? Which are the possible causes that can generate this error?
Thanks,
itelcom
7 Posts
0
May 20th, 2014 08:00
Hello,
yesterday we had another crash. We could manage to get the status before the device got stuck. Please, tell how can I attach a screenshot. Basically the result of a "sh process cpu" command was the voipTask process over 80%.
itelcom
7 Posts
0
May 20th, 2014 09:00
ok. Here it is.
itelcom
7 Posts
0
May 21st, 2014 01:00
Sorry, didn't update yet.
There are no IP phones plugged directly into it. There is an 8 x 3548 switches access stack, with all the ports in trunk mode, and near 300 IP h.323 phones plugged into.
We are thinking about erasing the "switchport voice autodetect auto" configuration of the 6224 stack, cause the packets are already tagged at the access stack, although we keep this configuration at other corporate sites without issues. What do you think?