Question MDD Timeout Errors ?

[mountainhouse]

After not seeing any MDD Timeout messages for 11 years, in mid April I started seeing on average 1 per day. Had Spectrum field tech out - he changed all the connections from the street to the modem. Problem did not go away. 2nd tech came out and replaced the drop from the street. Problem did not go away. Although I had no splitters I did have a coupler (1st tech relced it) so I moved the modem to my garage so that there was only a 8 foot RG6 coax from the grounding block where the drop form the street terminated. Problem did not go away. Bought a new modem. Problem did not go away.

Downstream power levels are in the 3 - 6.5 dBmV range and SNR 40.9 - 41.4 range. Upstream power 43 - 44 range. Same on the old and new modem. When the 1st tech was out he removed a forward antenuator that Spectrum had installed a few years ago as he thought that upstream power levels were too high (50 - 51.5)

Have a Field Tech coming out again - seems that it clearly is not my modem or my wriing that is causing the MDD Timeouts - any ideas? I've been told by the two techs that they could be due to a loose connection but everything is tight and all new coax from the tap on the street. Did see a posting that was similar where the issue was in the Node. Any ideas?

The new modem is DOSCIS 3.1 which I am new too. Netgear. I see 31 locked downstream channels with channel 25 not shown, but lower where it shows OFDM channels it lists channel 25 as locked with power of 7.3 dBvM. It also shows millions and increasing numbers of Unerrored Codewords and Correctable Codewords - about the same number for each. No Uncorrectable Codewords From other postings of the signal levels I am seeing am I correct that the 32nd channel is used for CFDM? But can anybody explain what the Codewords are all about?

Thanks for any assistance.

 

[bill001g]

Do you have a actual issue. Does you see packet loss and/or does the connection go down at times.

The MDD message is rated at warning level on my modem and that generally means it is not a issue by itself. If you have other message there maybe a larger issue.

They key thing on the codeword stuff is you have very low numbers of "uncorrectable" code words. This indicates data that was damaged so badly it was discarded. The other 2 are packets that had no errors and packets that had errors that the modem could fix. There are extra bits transmitted in the packets to allow a single bit here and there to be fixed but it can't generally fix multibit data loss.

 

[mountainhouse]

Do you have a actual issue. Does you see packet loss and/or does the connection go down at times.

The MDD message is rated at warning level on my modem and that generally means it is not a issue by itself. If you have other message there maybe a larger issue.

They key thing on the codeword stuff is you have very low numbers of "uncorrectable" code words. This indicates data that was damaged so badly it was discarded. The other 2 are packets that had no errors and packets that had errors that the modem could fix. There are extra bits transmitted in the packets to allow a single bit here and there to be fixed but it can't generally fix multibit data loss.

Did have a drop in service and it took several reboots of the modem to reestablish service on May 7th, while the neighbor on the same tap stayed up.

Other than that I am not aware of any drops, but with 11 years of never having had an MDD error at this location, then to have them once a day fro the last month to me is a sign that something has changed and could be the "canary singing in the coal mine" In twenty years of cable modems I had never seen an MDD error which is why this really sticks out to me.

 

[bill001g]

This is where you have to be a guy who works for the cable company and can see the messages on the ISP side. Most these messages only mean something when you know how the other end is configured. Many cable modem actually send these messages to the ISP equipment so they can in effect see your logs.

Not sure been a long times since I look at all the docsis messages. I just know the bad ones like sync failures.

 

[mountainhouse]

Have talked to 4 tech support reps - first one said he couldn't see the MDD errors but when pressed he asked for help and was shown how to see them, but he and his support had no idea what they are, Next 3 all could see them but also had no idea what they are either. First Field tech had no idea either and called another field tech who told him a loose connector so he redid them all. 2nd field (sho may have been the one the first field tech called as their are only 4 in that office) replaced the drop and checked all connectors - checked for ingress and finding a little on the in house coax (no splitters) wanted me to replace the in house coax, so before doing that I moved the modem right to where the drop enters the house and still have errors. New modem and coax (RG11) and still have errors. Field tech coming out again today.

 

[Ralston18]

Just as a matter of curiousity:

Does the MDD Timeout message or any other modem log entries include a MAC? If so, can you identify that device?

= = = =

Once a day, starting in mid-April....

Pattern? What time of day?

Take a look in Reliability History/Monitor and Event Viewer.

Are there any error codes, warnings, or even informational events that match that pattern?

Start with Reliability History - much more end user friendly and the time line format may prove helpful.

Event viewer requires more time and effort to navigate and understand.

FYI to help -

How To - How to use Windows 10 Event Viewer | Tom's Hardware Forum (tomshardware.com)

Also look in Update History for any failed or problem updates. Mid-april.

 

[mountainhouse]

Just as a matter of curiousity:

Does the MDD Timeout message or any other modem log entries include a MAC? If so, can you identify that device?

= = = =

Once a day, starting in mid-April....

Pattern? What time of day?

Take a look in Reliability History/Monitor and Event Viewer.

Are there any error codes, warnings, or even informational events that match that pattern?

Start with Reliability History - much more end user friendly and the time line format may prove helpful.

Event viewer requires more time and effort to navigate and understand.

FYI to help -

How To - How to use Windows 10 Event Viewer | Tom's Hardware Forum (tomshardware.com)

Also look in Update History for any failed or problem updates. Mid-april.

Yes - the MAC address in the Event log is the MAC address of the modem. The issue is between the modem and the plant.

No set pattern or time of day when it occurs - most days one error logged somedays no errors followed by two errors the next day.

On the old modem Arris SB6190) on some days would see a T3 Timout or Partial service 2 - 3+ hours later. On the new modem (Netgear CM2000), I have only had two MDD errors but both had a "CM-STATUS message sent. Event Type Code: 4;" sent 10 seconds later. Most likely just a difference in the way the different modems log things.

When this was first noticed all laptops had been powered off for 4 weeks and the MDD errors started after they were powered off and were logged on average once a day even though computers were all power off.

When the new modem was provissioned yesterday my router was disconnected as I was plugged directly into the modem with my laptop and was in that config when the two errors were logged in the modems' event log.

Received a call from the Field Engineer supervisor this morning. He had been researching the issue and agreed that it is nothing on my end - basically eveything has been removed to a direct connection and the new modem. Something in the plant end.

He had reached out already to the mainteance group (group that supports the plant - tap, node, etc.) and they told him the line needed to be balanced. He wasn't sure what that meant but I do see that the OFDM channel is about 1 db lower right now that it was this morning and the the down stream channels power levels are in a tighter range.

At this point we are going to sit tight and see if the balancing resolved the issue or not and will touch base in a few days.
The only other item that points to the plant is the the neighbor who is on the same tap (just the two of us) apparantly has a "crap load" of errors on her line. So tjhe supervisor will be reviewing that as well.

Just as a matter of curiousity:

Does the MDD Timeout message or any other modem log entries include a MAC? If so, can you identify that device?

= = = =

Once a day, starting in mid-April....

Pattern? What time of day?

Take a look in Reliability History/Monitor and Event Viewer.

Are there any error codes, warnings, or even informational events that match that pattern?

Start with Reliability History - much more end user friendly and the time line format may prove helpful.

Event viewer requires more time and effort to navigate and understand.

FYI to help -

How To - How to use Windows 10 Event Viewer | Tom's Hardware Forum (tomshardware.com)

Also look in Update History for any failed or problem updates. Mid-april.

MAC address is the address of the modem. Issue is between the modem and the plant. Errors are logged even when all computers are powered off. No pattern as to when it occurs other than most days have one error logged, some days have zero but others have two. While gone for the past 6 days, ahd 8 errors logged.

MDD messages are sent every couple of seconds to the modem and are logged as a timeout when not received is my understanding at a high level.

Not a computer related issue as when the modem was provissioned yesterday it was connected directly to the modem and two errors were logged while in this config - nothing reported in stability history or event viewer for the time frame of the errors. And as I said - when this first occured all computers were powered off for 4 weeks.

Field tech supervisor escalated to the maintenance group this morning who told him the line needed to be balanced. Noticed afterwards that the power level on the OFDM channel is about 1 db lower than this morning and that the power levels on teh downstream channels seem to be tighter. Going to watch logs for a few days and see if that fixed it. In the meantime he is working with the maintenace group to have them check the tap and node, as the last field tech reported that the neighbor who is the only other one on the tap had a crap load of errors - he just wouldn't provide me with more info - privacy and all.

 

[mountainhouse]

Update - Field Supervisor has taken a deep look at the node I am on and seeing low SNR on upstream channels - higher than average T3 timeouts as well - two x 3 times average for other areas. Still in the process of escalating to the maintenance team.