PCIe TE121 card causing server to freeze

Asterisk Asterisk Support
1

Hi - Does anybody have issue in installing a TE121 (PCIe version) in Dell Poweredge SC440 and Dell Optiplex 755?

I am having a problem where the server freeze after a day or so with no error indication in any of the system logs, even the mobitor screen is blank.

I was suspecting it could be the defected TE121 card. However it cannot be the case sincei have 4 of these machines and all of them have the same issue.

Thanks
Foong

2

Foong,

Thanks for giving Digium’s Technical Support team (available to any Digium customer at support@digium.com or +1 256 428 6080) an opportunity to address this issue. I see in our internal ticketing system that we are currently working on this issue. We hope to have a resolution for you soon.

Regards,
~Kenny Shumard
Digium Technical Support Manager

3

Hi,

We have the similar issue on our Dell server 1950 and 2950.

Should I return the card to the reseller?

Thanks

4

hellbound,

Please contact Digium Support at support@digium.com or +1 256 428 6161 to troubleshoot the issue you’re experiencing. I wouldn’t recommend taking the card back to the reseller until we’ve had a chance to identify the problem – it’s possible that we can get the current card working without having to replace it.

Thanks!
~Kenny Shumard
Digium Technical Support Manager

5

Hi all,

I have the same issue on a server with the following configuration:

Mb ASUS P5E-VM DO Socket 775
CPU Intel Core 2 Duo E7200 2,53GHz, 3Mb, 1066MHz Socket-775 BOX
2xDIMM 1024Mb PC2-8500(1066Mhz) Kingston (KVR1066D2N7K2/2G)

So, the problem with this card is not apparently restricted with Dell servers only…

Are there any updates on this issue? Will there be a solution and when? Or shall we replace the card?

Thank you,
Dmitry

6

we had no choice to change the card to sangoma,

but I will interested to know how to fix it.

thanks

7

Did you contact support? If not, you really should they do a great job. If so what did they say?

8

Yes, I’ve sent them an email and got an auto-reply confirming that my ticket has been created. Now I’m waiting for an answer, hopefully with a solution.
Has anyone else used TE121? What was the experience? I still don’t know what the root of the problem is: is my particular card buggy, is it a common problem with all TE121 cards or is the card incompatible with my server config. Any useful information will be much appreciated.

9

I’ve been waiting for more than two weeks before I get a initial respond. they’ve asked a few question and still waiting for final report on this.

10

[quote=“hellbound”]Hi,

We have the similar issue on our Dell server 1950 and 2950.

Should I return the card to the reseller?

Thanks[/quote]

Hellbound, Im having the same issue with one computer here…we 've bought one supermicro motherboard x7dvl-e…and its happening the same thing…I think it looks like irq conflict…

-Thiago

11

I bought 2 TE121B’s a few weeks ago from e4 and tracked this thread because I was leery of whether I’d have problems with them. I have one running in a Dell Poweredge 850 and one in a Dell Precision 690 for two weeks with absolutely no problems. I have a PRI from AT&T and the other one from XO. I’m using Elastix 1.2.1. Here’s the details from the sticker on the back of my card, maybe you have a buggy revision.
Part #5TE121 LF
Rev. B WO#9245

12

Have logged a support call with digium - no resolution yet, seems like they are aware of the problem after speaking to them on the phone (although they did not have a workaround or resolution).

I was told to try removing the echo cancellation module, however this did not fix the problem.

The server will be running fine for a period of time before locking up entirely (no network connectivity, frozen terminal window).

I noticed the following log messages in /var/log/messages just prior to the freeze:

…
Oct 13 11:56:33 msb kernel: wcte12xp0: Missed interrupt. Increasing latency to 4 ms in order to compensate.
Oct 13 11:57:30 msb kernel: wcte12xp0: Missed interrupt. Increasing latency to 5 ms in order to compensate.
Oct 13 11:57:44 msb kernel: wcte12xp0: Missed interrupt. Increasing latency to 6 ms in order to compensate.
…

I have tested on different servers using different linux distributions (both RHEL5 and Ubuntu Hardy) - all of which experience the same lock ups. I have also tested with rebuilt kernels (running 2.6.27).

I have compiled and installed the latest asterisk (1.4.22) , dahdi (linux/tools 2.0.0) and lib-pri (1.4.7) software.

I’ve sent them the requested diagnostic information, hopefully their engineers can replicate and solve the problem ASAP.

13

I got the problem!!!for some reason with PCI-e was giving me the error…I got one PCI normal, and it is working FINE!!!I hope if you have the chance to try with one PCI (not pci-e) maybe help you out!

-Thiago

[quote=“gavz”]Have logged a support call with digium - no resolution yet,
seems like they are aware of the problem after speaking to them on the phone (although they did not have a workaround or resolution).

I was told to try removing the echo cancellation module, however this did not fix the problem.

The server will be running fine for a period of time before locking up entirely (no network connectivity, frozen terminal window).

I noticed the following log messages in /var/log/messages just prior to the freeze:

…
Oct 13 11:56:33 msb kernel: wcte12xp0: Missed interrupt. Increasing latency to 4 ms in order to compensate.
Oct 13 11:57:30 msb kernel: wcte12xp0: Missed interrupt. Increasing latency to 5 ms in order to compensate.
Oct 13 11:57:44 msb kernel: wcte12xp0: Missed interrupt. Increasing latency to 6 ms in order to compensate.
…

I have tested on different servers using different linux distributions (both RHEL5 and Ubuntu Hardy) - all of which experience the same lock ups. I have also tested with rebuilt kernels (running 2.6.27).

I have compiled and installed the latest asterisk (1.4.22) , dahdi (linux/tools 2.0.0) and lib-pri (1.4.7) software.

I’ve sent them the requested diagnostic information, hopefully their engineers can replicate and solve the problem ASAP.[/quote]

14

I am currently experiencing the same problem with an AEX808E card. I have also tested multiple distributions (ubuntu, centos, fedora), and two different motherboards (intel MBPDG31PR & intel D945GCNL).

Is this an issue that can be resolved, or is it hardware failure on the card itself?

15

Hello,

i have same problem…

First i had a small server with old TE100 (i think) PRI PCI card conf E1, CELERON prozessor. Works fine for a couple of hours. Hangs!

So i thougt maybe something wrong with the Hardware an bought a complete new System - iE Athlon X2, PCIe TE121 PRI card. It then worked fine for about 8 hours. CRASH. 4 hours. CRASH. (PS: PBX in a FLASH, Asterisk 1.4)

Other ASTERISK without PRI Card working fine for Years now!

So: It seems it ist not a special kind of HARDWARE (PCI vs. PCIe, completly diffrent Mainboards (INTEL Low-Cost vs. ASUS Athlon)

I am not sure because i am not LINUX specialist but i think it could only be some KERNEL Problem in combination with lots of IRQs.

Also i think the faster Athlon seems to run a little longer befor CRASH.
I can see that on day (less or 0 incomming calls on E1 PRI Card) the system seems to work. At night when the useres beginn to use the phones after work it crashes very fast.

16

Digium has no “TE100” product. You may be thinking of the T100P, the E100P, or the TE110P, all of which are end-of-life cards and no longer supported. At Digium, we stand behind the quality of our cards, and have been able to work through some of the issues reported in this thread.

I think this thread has become a common place to share “my card failed, too!” stories, even though several of the problems reported are unrelated – sometimes even for different cards. I’d like to reiterate that Digium Support is here to help, and in almost every case we can find a satisfactory resolution if given a chance to do so. This forum is full of knowledge about Asterisk and Digium products, but you’ll always be able to get the most up-to-date information by working directly with our Support team.

As always, we’re available at support@digium.com or +1 256 428 6161. Please, let us help you directly! We don’t monitor every thread in the forum!

17
  1. Sure you are right, the product is E100p, thats why i wrote ‘i think’.
  2. the new TE121 is better (EchoCancel) so i would like to get thisone working if possible. I took hours in trying out. The problem is that the “hang” is not coming fast. But: I use PBX in a FLASH, and after an UPDATE-SOURCE to the newest source-es incl. KERNEL the ERROR is reproducable within about 30 min. ALLWAYS the Screen is then going to be scrambled. Everything is frozen.
  3. Exactly the same Hardware/Software but without the E100P works in 3 other locations without any problem.
  4. I now try out an other mainboard whith CELERON, 2GB RAM… we will see.

PS: Thanks for the reply! As a customer i am VERY happy with ASTERISK and i surly believe that your Hardware is made to good standards - but i also know that it is hard for you to maintain all combinations of Mainboard and so on.

18

This won’t make you feel any better, but…

I had constant problems with a TE121 PCIe. Numerous support cases with Digium. Sometimes the TE121 would work for a few days at a time - mostly, it wouldn’t hold a solid T1 for more than 30 minutes. I tried every troubleshooting step on every website - on its own irq, not running X, IDE DMA mode, etc on 4 different machines. Nothing solved the problem.

I bought a TE412P PCI 3.3v.
It’s been solid as a rock from the second I put it in the server.
Works great on a Dell PowerEdge 2800 and a Dell PowerEdge 1950 III.

I don’t know if the solution was to get away from the PCIe or the single-span card, but I wouldn’t touch one of those again!

Good luck to you.

19

Hello “jcromes”,

this sounds good. I ordered the TE412P (TE410P + sep. EchoCancel Module) now. My dealer does not have the bundle in stock, so i pay bit more.
I will tell the community on monday if it works.

Thanks a lot for that hint!

And:
Seems to be the single-span not the PCIe. Because i had same problems with E100p which is not PCIe. But: no experiments now. My useres are so peeved - i cant try out!

And to “kshumard”:
Seems DIGIUM must analyse the problem:
Try out:
ASUS M2N-MX SE PLUS with AM2 Athlon X2, SATA Hard Disk, TE121, Booting from USB-CD installing PIAF 1.3 Asterisk 1.4.
Other side of TE121: E1 (span=1,1,0,ccs,hdb3,crc4)
PS: On ASUS P5N73-AM with CELERON an sep. REALTEK Ethernet-Card PIAF (PBX in a FLASH) SATA, you can make LOOPBACK-TEST and
you can see immedeatly SCREEN beeing not OK. I cant test ASUS M2N now - i have to reinstall again.
So: Good luck to DIGIUM, hope they will find the problem.

20

Hello

I have the same problem, my server frozes after several messages in /var/log/messages :

…
kernel: wcte12xp: Wrote ‘fb’ but read 'fffffe00’
kernel: wcte12xp: Wrote ‘fc’ but read 'fffffe00’
kernel: wcte12xp: Wrote ‘fd’ but read 'fffffe00’
kernel: wcte12xp: Wrote ‘fe’ but read ‘fffffe00’
…
kernel: wcte12xp: Setting up global serial parameters for E1
kernel: wcte12xp: Found a Wildcard TE121
kernel: Uhhuh. NMI received for unknown reason 00 on CPU 0.
kernel: Do you have a strange power saving mode enabled?
kernel: Dazed and confused, but trying to continue
kernel: wcte12xp31: Missed interrupt. Increasing latency to 11 ms in order to compensate.
kernel: Uhhuh. NMI received for unknown reason 00 on CPU 0.
kernel: Do you have a strange power saving mode enabled?
kernel: Dazed and confused, but trying to continue
kernel: wcte12xp31: Missed interrupt. Increasing latency to 12 ms in order to compensate.
kernel: Uhhuh. NMI received for unknown reason 00 on CPU 0.
kernel: Do you have a strange power saving mode enabled?
kernel: Dazed and confused, but trying to continue
…

Config : dahdi 2.0.0, asterisk 1.6.0.1, libpri 1.4.7

Is there some patch yet ?
Regards