Corrected blocks going through the roof! Page 2
Links: home · search · speed test · login · more ·
Links: Reply New Topic
Forums » AT&T » AT&T U-verse » Corrected blocks going through the roof!
page: 1 · 2 · 3
gdm @ 4th Nov 11:27PM:
Re: Corrected blocks going through the roof!
Just because you don't see it on the TV doesn't mean anything. Corrected blocks are normal and some have hardly any and others have alot. What defines alot not sure I can tell you that either but in a 24 hour period 100,000 is lot.
When I have heavy rain I get anywhere from 15k to 40k in 15 minutes. It does causes issues with the TV only voice and internet I have no issues.
The thing to remember is with U-Verse we are all setup differently to some extent internally and even externally. Different gauge wire on the poles, aerial vs. underground, etc.
reply
Keoway @ 4th Nov 11:30PM:
Re: Corrected blocks going through the roof!
I'll reboot again, if it produces runs ;-)
Here's what I'm seeing after about 2 hours of up-time:
Link Retrains: 5 5 0 1:52:36
DSL Training Errors: 2 2 0 1:51:49
Training Timeouts: 2 2 0 1:55:48
Loss of Framing Failures: 0 0 0 0:00:00
Loss of Signal Failures: 0 0 0 0:00:00
Loss of Margin Failures: 0 0 0 0:00:00
Cumulative Seconds w/Errors: 14 14 0 0:35:11
Cumulative Sec. w/Severe Errors: 0 0 0 0:00:00
Corrected Blocks: 59938 59938 0 0:17:33
Uncorrectable Blocks: 21 21 0 0:35:11
DSL Unavailable Seconds: 284 284 0 1:51:51
reply
Tanshin @ 4th Nov 11:34PM:
Re: Corrected blocks going through the roof!
It seems to be slowing (especially since it hasn't gone up in 17 minutes). But we'll have to see what happens over longer time.
What is the weather situation down where you are? How about where the lines are? Just kinda curious...
reply
Keoway @ 4th Nov 11:51PM:
Re: Corrected blocks going through the roof!
There was heavy rain for most of Saturday. Temperatures continue to fluctuate. On Monday and Tuesday, those were the somewhat odd type of days, where the AC needs to be active for when one returns home in the evening, but then the thermostat needs to be set to Heat for the morning. Right now, its 51 with 81% humidity. Tomorrow, there will be around a 20 degree delta between the morning low of 46 and projected high of 66.
Most of my line, if not all, is underground to the MDU's Ped. The associated crossbox, DSLAM, and VRAD are in one of those roadside easement locations that's backfilled with gravel and in a heavily shaded area.
reply
anon @ 5th Nov 01:08AM:
Re: Corrected blocks going through the roof!
Seeing Corrected blocks incrementing means impulse noise protection is working, whatever line problems you have, aren't bad enough that they can't be corrected with Reed Solomon FEC and interleaving. Corrected blocks just means the parity bytes from the RS codeword are being used to recover the DMT symbol.
Uncorrected blocks simply mean the DMT symbol can't be recovered and is dropped, hence the upper layer PDU (PTM frame in this case) will be dropped as well. One uncorrected block is at least one dropped Ethernet frame.
reply
ccastiglia @ 5th Nov 09:54AM:
Re: Corrected blocks going through the roof!
I am just going to add that I have had the same problem with corrected blocks on several occasions (I've posted about it here and/or in AT&T Direct forum in the past).
The one item I want to add to the conversation is this: my RG would go for days or weeks without a single, or at most a few, corrected blocks. Then, as if something went terribly wrong, suddenly every new block was being corrected with the numbers quickly reaching the millions, hundreds of millions, and billions until I rebooted the RG. After the reboot, things would go back to normal for a few weeks until the next episode began.
And yes, during these episodes I would experience TV frame freezes and pixellation events.
reply
anon @ 5th Nov 11:23AM:
Re: Corrected blocks going through the roof!
I find it interesting that the current version is 5.29.135.47.
I'm in North Atlanta, and my RG is still using 5.29.105.94. I wonder when we'll get the latest and greatest. update.
For what it's worth, I rebooteed two days ago, and here are my numbers as of a minute ago.
DSL Corrected Blocks: 121899 1959 192 0:06:13
reply
ccastiglia @ 5th Nov 11:38AM:
Re: Corrected blocks going through the roof!
This was my post in AT&T Direct forum last May
-------------------------------
DSL Corrected Blocks
My U-Verse connection consistently racks up humongus numbers of "DSL Corrected Blocks". Is this normal?
IP Bytes Packets Errors
Transmit: 468610811 2734306 0
Receive: 557586885 73888736 0
Data Errors
Statistics
Collected for 11 days 11:35:24
Since Reset Current 24-Hour Interval Current 15-Minute Interval Time Since Last Event
DSL Link Retrains: 0 0 0 0:00:00
DSL Training Errors: 1 0 0 11 days 11:33:09
DSL Training Timeouts: 2 0 0 11 days 11:33:32
DSL Loss of Framing Failures: 0 0 0 0:00:00
DSL Loss of Signal Failures: 0 0 0 0:00:00
DSL Loss of Power Failures: 0 0 0 0:00:00
DSL Loss of Margin Failures: 0 0 0 0:00:00
DSL Cumulative Errored Seconds: 46 0 0 1 day 11:48:38
DSL Severely Errored Seconds: 4 0 0 1 day 11:59:31
DSL Corrected Blocks: 2590858544 145964180 1129519 0:00:01
DSL Uncorrected Blocks: 279 0 0 1 day 11:48:38
reply
aturner733 @ 5th Nov 12:52PM:
Re: Corrected blocks going through the roof!
said by ccastiglia :
I am just going to add that I have had the same problem with corrected blocks on several occasions (I've posted about it here and/or in AT&T Direct forum in the past).
The one item I want to add to the conversation is this: my RG would go for days or weeks without a single, or at most a few, corrected blocks. Then, as if something went terribly wrong, suddenly every new block was being corrected with the numbers quickly reaching the millions, hundreds of millions, and billions until I rebooted the RG. After the reboot, things would go back to normal for a few weeks until the next episode began.
And yes, during these episodes I would experience TV frame freezes and pixellation events.
Glad to hear that I am not the only one with this same issue. Mine does the same thing, ok for a while after a reboot, then goes crazy.
reply
marrow @ 5th Nov 01:07PM:
Re: Corrected blocks going through the roof!
Here is what was done over the course of a month and a half to fix this problem:
1. Re-grounded the NID (box outside your house where the drop connects) (AT&T field tech)
2. Replaced the HISA Balum (box inside the NID) (AT&T field tech)
3. Replaced the line drop from the pole to the NID (AT&T I/R tech)
4. Replaced the line from the NID to the gateway X 2 (I have coax, phone line, and cat 5 that can be and have been used to go into the gateway). (AT&T field tech)
5. Replaced the Gateway X 2 (AT&T field tech)
6. New Ground for the gateway plug (Electrician)
7. Replace Ground for the NID (AT&T field tech and I/R tech)
8. Replaced two of the protectors (plastic piece with two screws on the top) within the NID (I think this is what fixed the problem as the other fixes showed minimal improvements) (AT&T field tech)
If someone is having a similar problem and the suggestions above do not help, have the field tech test the line at the drop. If he/she can see the problem there, then they need to send an I/R tech (line tech). If the problem is not there, have the field tech test on the other side of the NID and on down the line until they find where the problem is occurring.
reply
Keoway @ 5th Nov 11:29PM:
Re: Corrected blocks going through the roof!
An electrical ground issue is also something that I've encountered. When rearranging my TV's placement and clutter collection for easier access by the Prem Tech, I discovered that the intended outlet had an open ground.
While waiting for it to be serviced/repaired, I have had to resort to use of a 10ft(?) extension cord. It directly feeds a Kingston power strip where Cisco, 2Wire, Belkin, Samsung, and JVC all cohabit. [The other half of the outlet has a Belkin strip which supports a lamp that has a CFL.]
Here's what I have this evening, after last night's reboot:
Cumulative Seconds w/Errors: 87 4 0 0:09:55
Cumulative Sec. w/Severe Errors: 0 0 0 0:00:00
Corrected Blocks: 800325 5517 776 0:00:00
reply
btur389w @ 6th Nov 03:13PM:
Re: Corrected blocks going through the roof!
have you tried removing the Lamp and CFL from that circuit? some CFL Ballasts cause a small amount of Feedback that resonnates through the circuit. usually a circuit breaker takes care of resonnace for you, but if there isn't a circuit breaker between the RG and the ballast, it could cause an issue. please let me know if that helps out with those numbers any.
--
If you have any questions or concerns, please feel free to email us at uversecare@att.com.
Our goal is to provide you Excellent Customer Service. Hours: Tu-Sa 12-9 CST
reply
Keoway @ 6th Nov 06:36PM:
Re: Corrected blocks going through the roof!
I use the CFL all the time, but some days it appears to be having a "bad" day and flickers. I'll try and notice whether there is any correlation with the skewed periods of corrected blocks. I'll also monitor periods when the lamp/CFL has been removed from the circuit.
With all of these discussions about various environmental issues, it also makes me wonder whether the extension cord, snaked across the floor and providing interim grounded outlet support, has any potential causality. As a 2nd floor dweller, in an MDU environment, the opportunity's for EMI are rampant. The structural elements for common walls/floors probably don't provide much shielding from EMI opportunities that my neighbors might present.
Here's what I' have this evening:
Cumulative Seconds w/Errors: 120 37 0 1:13:22
Corrected Blocks: 1229935 435127 0 0:30:05
Uncorrectable Blocks: 229 67 0 1:13:22
reply
nwrickert @ 7th Nov 06:03PM:
Re: Corrected blocks going through the roof!
They seem to be relentless and don't seem to be stopping, but so far the video looks fine this evening.
I started seeing the same "problem". Corrected blocks are accumulating at a high rate.
I am suspecting a bug in the RG firmware. At present there is no TV on, and no heavy downloading. A few minutes ago, I did a reset of the statistics so that they are counting from zero. And I am seeing:
DSL Corrected Blocks: 380668 380668 380668 0:00:00
Given the relatively low level of activity, that must be counting all blocks as requested. Or the count may even be greater than the number of blocks it has handled.
Oops, the count is now up to:
DSL Corrected Blocks: 557498 557498 557498 0:00:00
--
AT&T Uverse; Zyxel NBG334W router (behind the 2wire gateway); openSuSE 11.0; firefox 3.0.15
reply
Keoway @ 7th Nov 06:29PM:
Re: Corrected blocks going through the roof!
After some new tests, I'm beginning to fixate on EMI within the electrical circuit where the RG is hosted and "noise" from a CFL.
I recently discovered a second deficient circuit* that has an "Open Ground" state. This evening, I added another extension cord to my collection of wiring spaghetti, in order to move the CFL to a properly grounded circuit that is different than where the RG is hosted.
I'll be monitoring closely during tonight's CU and FSU game, so I'll have new numbers later this weekend.
* To check the condition of a receptacle, I'm using one of these GB Circuit/Receptacle testers.
»www.filmtools.com/idcirtes61.html
reply
btur389w @ 7th Nov 06:34PM:
Re: Corrected blocks going through the roof!
our techs have a very simlure tester to that. it is accurate for your tests.
reply
Keoway @ 7th Nov 06:42PM:
Re: Corrected blocks going through the roof!
said by btur389w :
our techs have a very simlure tester to that. it is accurate for your tests.
Our email sidebar discussion, marrow's remarks about his circuit issue, and ACE Hardware's 20% "Fill a Bag" sale made me go shopping today! I have new toys ;-)
reply
btur389w @ 7th Nov 06:49PM:
Re: Corrected blocks going through the roof!
Haha, sorry to have given you the compulsion to shop for toys. my wife won't let me shop :(
let me know how it goes.
reply
Tanshin @ 7th Nov 09:01PM:
Re: Corrected blocks going through the roof!
Shoot, I forgot it was the Bag Day. They usually send a little "bag" in the mail to remind you... but they didn't this year.
Oh well!
reply
Keoway @ 8th Nov 12:48AM:
Re: Corrected blocks going through the roof!
I reset the RG stats at 18:00, here's the non-zero data 6.8 hours later:
Cumulative Seconds w/Errors: 40 40 0 1:43:42
Corrected Blocks: 96773 96773 0 0:11:06
Uncorrectable Blocks: 70 70 0 1:43:42
This greatly improved results is after removing the CFL from Grounded_Circuit #1. The RG resides on the circuit.
The CFL has been moved to Grounded_Circuit #2. My first attempt to relocate the CFL encountered a second deficient circuit. When the CFL was "visiting" on Improperly_Grounded_Circuit #2, the CFL still seemed to be throwing "noise" back to the RG.
reply
Tanshin @ 8th Nov 10:29AM:
Re: Corrected blocks going through the roof!
Wow, that's amazing how that can cause those kinds of problems.
reply
nwrickert @ 8th Nov 10:47AM:
Re: Corrected blocks going through the roof!
I started seeing the same "problem". Corrected blocks are accumulating at a high rate.
I rebooted the RG this morning. The rate of reported corrected blocks is now sane.
It takes a long time to properly synch after reboot.
--
AT&T Uverse; Zyxel NBG334W router (behind the 2wire gateway); openSuSE 11.0; firefox 3.0.15
reply
Keoway @ 8th Nov 10:49AM:
Re: Corrected blocks going through the roof!
Time to start a new day. Here's the numbers from early this morning:
Cumulative Seconds w/Errors: 48 48 0 0:40:56
Corrected Blocks: 285423 285423 836 0:01:50
Uncorrectable Blocks: 94 94 0 0:40:56
I booted the laptop, before touching anything else that needed power and checked the stats. Corrected block count had been 0 for 35 minutes, but then I turned on the CFL. Just turning it on ... the count went to +400. I cycled it and the count went to 836.
No hits occurred when I powered up the TV, STB, and tuner/receiver.
reply
bsnelson @ 8th Nov 05:52PM:
Re: Corrected blocks going through the roof!
So I guess these numbers are a little on the high side?
Collected for 0:47:53
DSL Since Reset Current 24-Hour Interval Current 15-Minute Interval Time Since
Last Event
Link Retrains: 0 0 0 0:00:00
DSL Training Errors: 0 0 0 0:00:00
Training Timeouts: 0 0 0 0:00:00
Loss of Framing Failures: 0 0 0 0:00:00
Loss of Signal Failures: 0 0 0 0:00:00
Loss of Margin Failures: 0 0 0 0:00:00
Cumulative Seconds w/Errors: 1 1 0 0:06:53
Cumulative Sec. w/Severe Errors: 0 0 0 0:00:00
Corrected Blocks: 24785791 24785791 1500952 0:00:00
Uncorrectable Blocks: 2 2 0 0:06:53
DSL Unavailable Seconds: 0 0 0 0:00:00
Almost 25 million corrected blocks in a little over 45 minutes? Few uncorrected, though. I'm not seeing any obvious performance problems, but I'm an internet-only customer and am not really pegging the link by any means.
Is this something I should be concerned about? I'm still on the "old" software and I haven't rebooted the RG in probably a month or more (although I obviously reset the statistics for this snapshot).
Brad
reply
Keoway @ 8th Nov 11:05PM:
Re: Corrected blocks going through the roof!
Here's 28 hours of stats, after relocation of the CFL to a different, grounded, circuit:
Cumulative Seconds w/Errors: 166 85 0 1:22:53
Corrected Blocks: 777532 304297 75 0:01:03
Uncorrectable Blocks: 273 136 0 1:22:53
I like the way these numbers look! This U-Verse diagnostic voyage has also made me reflect upon what had become a huge annoyance with my home network.
Prior to switching to U-Verse, I had FastAccess for more than 8 years. I ran the original Alcatel STH modem the entire time with initially a D-Link 802.11b AP and later a Netgear 802.11g AP.
For several months, I had started to experience lots of "drops" from the Netgear box. I was starting to question its health and/or that of my laptop's wireless radio. I also wondered about "noise" from all of the other SSID's that wonder around the neighborhood.
The "drops" always seemed to be minimal when I was on battery power. When on AC power, the incident rate usually seemed to climb. The laptop charger, CFL/lamp, HP printer, and shredder all resided on the same Belkin power strip.
I may have wasted lots of time and got really disgusted by the earlier problem ... when the solution was real simple --- Unlpug the lamp! (The CFL in question is a 6500K 23W bulb.)
Nota Bene: Individual mileage may vary. Objects in mirror are closer than they appear ...
reply
Tanshin @ 9th Nov 03:13PM:
Re: Corrected blocks going through the roof!
This is just so interesting. I really am astounded by some of these things!
reply
bsnelson @ 10th Nov 08:45AM:
Re: Corrected blocks going through the roof!
As another poster suggested, I rebooted my RG last night, and have only had 2,781 corrected blocks in the almost 10 hours since. I was getting that many per SECOND before.
I'm thinking bug in the statistics gathering code at this point.
Brad
reply
anon @ 11th Nov 10:24PM:
Re: Corrected blocks going through the roof!
Corrected Blocked are not a bad thing. Uncorrected blocks are. When corrected blocks happens, and it works. That means the equipment is doing its job.
Unless this is service impacting, do not worry about it.
Generally when it is service impacting, there is another issue that aggravates this. Usually, though, it is not.
reply
doublea @ 13th Nov 02:30AM:
Re: Corrected blocks going through the roof!
It's interesting to read through all of these pages about corrected and uncorrected blocks. As some of you may have read through my story with corrected and uncorrected blocks, I had Matt help me, we ended up having our entire walkway torn up, and the techs looking for our line.... We had a guy check all of the connections of our line all the way to the vrad, and "clean" the line. We had crews out here several times. I have on average 4 billion corrected blocks in a day, and 1 million uncorrected.
Now I don't really notice a performance drop other than getting a page cant be displayed messages a few times a day. Ive fully gotten used to it.... why?
It turns out the crews wanted to tear up the road in front of my house. Normally this wouldn't be a problem, but this is a townhouse development. Just last year we put a few hundred thousand into redoing all of our roads perfectly (cost was split 72 ways, but still it was expensive), we don't want any patchwork on our new pavement. Or at least my neighbors don't, and I don't blame them. I never personally told AT&T to stop working on my problem, but I watched several of my neighbors yell the AT&T crews about the marking on the pavement, (markings prepping the areas to be jack hammered up. )
I'm glad to hear this possibly has something to do with the RG itself, My uncorrected, and corrected also stay very very low several days after a reboot, maybe 50 in 10 hours, with 0 uncorrected. So a RG reboot helps me too. So I admit, their is probably something with my line also, but I hope that a software update could at least fix some of the problem.
I'm also not willing to try get more help in fixing this, because I'm scared that AT&T will put me on a lower profile and call it fixed. I will live with my issues rather a slower speed.
Matt was very helpful, but my situation kinda left me in a place were what I have can't be fixed.
And no I don't want Comcast, they do have docsis3 here, but they cap, and we have a nasty taste left in our mouth left from a issue a few years ago with them. I'm still MUCH MUCH more happy with AT&T in general.
So short of tearing up my road, or lowering my speed, or a magical software patch for the RG, I continue to ignore my problems :P.
The attached image is from tonight.
reply
systech @ 22nd Nov 08:59PM:
Re: Corrected blocks going through the roof!
If indeed the outside line is good and RG is still showing mega corrected blocks there could be something else at work. We have encountered this a lot in the field. We have come to learn that the RGs are very sensitive to common mode and differential mode noise on the power line. There have been many cases where our test sets sync to perfect lines running well over max rates of 50Mbps while the RG will not sync at all or when they did the sync rate was terrible and running massive errors. Sometimes floating the ground breaks the ground noise loop and essentially fixes the issue. Other times we can install one of our new isolation transformers in line with either the RJ-11 phn cord or Coax depending on how the unit is fed. Either way you can isolate the issue by running the RG off of a battery while watching your diagnostics page. We have battery backups we use for this, but a regular 12V battery works just fine.
--
Americans don't care how much something costs, as long as they don't have to pay for it all at once.
reply
Keoway @ 22nd Nov 09:44PM:
Re: Corrected blocks going through the roof!
Here's my stats for 9 days 21:13:13, but reflect being away for 6 of the nearly 10 day period.
Cumulative Seconds w/Errors: 185 20 0 1:26:06
Corrected Blocks: 4,502,961 495,512 518 0:00:30
Uncorrectable Blocks: 442 61 0 1:26:06
When returning, I checked the stats and the cumulative corrected blocks count was circa 3,500,000. [Note: When away, I did use the iTouch applet to add/delete 3 DVR record events.]
When I have time to monitor the stats, I want to experiment with reintroducing the CFL back into the same circuit with the RG and observe whether a causality seems apparent.
Here's the Link stats to correlate with the error stats:
DSL Down Up
Current Rate: 25216 kbs 2048 kbs
Max Rate: 54496 kbs Not Available
Current Connection:
Current Noise Margin: 22.0 dB Not Available
Current Attenuation: 25.6 dB Not Available
Current Output Power: 13.6 dBm -7.3 dBm
IP Bytes Packets Errors %
Transmit: 508928136 2159643 0 0
Receive: 79480972935 59836200 7 0
reply
doublea @ 22nd Nov 11:19PM:
Re: Corrected blocks going through the roof!
said by systech :
If indeed the outside line is good and RG is still showing mega corrected blocks there could be something else at work. We have encountered this a lot in the field. We have come to learn that the RGs are very sensitive to common mode and differential mode noise on the power line. There have been many cases where our test sets sync to perfect lines running well over max rates of 50Mbps while the RG will not sync at all or when they did the sync rate was terrible and running massive errors. Sometimes floating the ground breaks the ground noise loop and essentially fixes the issue. Other times we can install one of our new isolation transformers in line with either the RJ-11 phn cord or Coax depending on how the unit is fed. Either way you can isolate the issue by running the RG off of a battery while watching your diagnostics page. We have battery backups we use for this, but a regular 12V battery works just fine.
This still does not explain the problem of when I reboot the RG all the errors, go away.... for a few days...
On the flip side, care to elaborate on the "ground" all the techs that came to my house said my line has a grounding issue, but im pretty much clueless on to what that means is up, and how they can fix that other than tearing my road up :P
Thanks for the reply.
reply
Thank you for using lo-fi dslreports.com - report bugs
© 99-2009 silver matrix LLC
