D-Link Forums
The Graveyard - Products No Longer Supported => Routers / COVR => DIR-655 => Topic started by: Johnnyboy! on June 27, 2009, 10:57:10 AM
-
So I *think* I have narrowed down my restart/instability issues down the NTP service on the router.
I'm using an A3 router with firmware 1.32b02 now, but I have had this problem with all firmwares 1.30+. The basic problem is that I have been trying to upload several gigabytes of files (not using http or ftp, but an encrypted file transfer over an obscure port). The problem occurs when randomly, my router restarts itself, thus cutting my file transfer and pissing me off (the program doesn't auto-resume).
I have noticed some odd issues with the NTP service. I have been using ntp1.dlink.com and time.nrc.ca ntp services. Both have the same issues.
The first issue, is that occasionally, after a restart (either intentional restart or not) the router time reverts back to the *exact* time that I initially flashed the firmware. If I have the NTP service enabled, it will eventually correct it, though I think some things get messed up before that.
Here's an older log file describing the problem:
[INFO] Fri Jun 26 07:17:37 2009 Time synchronized
[INFO] Sat Jan 31 11:49:44 2004 Requesting time from 207.232.83.70
[INFO] Sat Jan 31 11:49:44 2004 Time server ntp1.dlink.com is at IP address 207.232.83.70
[INFO] Sat Jan 31 11:49:44 2004 Next dynamic DNS update scheduled for Tue Feb 24 10:49:43 2004
[INFO] Sat Jan 31 11:49:44 2004 Stored configuration to non-volatile memory
[INFO] Sat Jan 31 11:49:43 2004 Successfully updated dynamic DNS entry for ***.no-ip.ca nochg ***
If you notice, after a restart, the time reverted back to Jan 31, 2004 (not sure the significance of this date) then changed the date back to Jun 26, 2009. But before it corrected the date it updated the DynDNS registry and set the next update for Feb 24th, 2004.
Would this cause a crash to have conflicting dates programmed into the scheduler? I don't know.
In the 1.31 release notes, it states 14. Router time is not reset when factory reset. But I guarantee that it does indeed reset itself.
I'm not sure where I am going with this, all I know is that something is screwy with the NTP feature of the router. I know other people have had lots of issues with this feature as well. Without the time service enable, my router has run for nearly 24hrs now without a restart, which is (sadly) amazing!
I'll keep poking on it, see if I can come up with the exact issue.
Any other ideas?
jB
-
Good work
I have also seen NTP issues as said in a previous thread. I can safely say that also pool.ntp.org gives problems so indeed it does not seem to matter which server you use.
For me the problem became visible with loading saved settings. Before upgrading the firmware i saved my settings (which are quite some settings like schedules, wifi, virtual server, etc.; incl NTP and other timing settings in that page)
When loading this file into the new firmware i can no longer log in to the router and it starts to reboot every time (login in will give the nice page "router is determining the speed of the network or something". But this page never goes away again.
i narrowed this problem down to the time setting page (where you can also set NTP). Leaving this page default results in being able to load your settings and login afterwards. (and the router does not keep rebooting).
Sadly these NTP problems seem not related to the overall stability. Not setting a NTP server will still result in the router not being able to grant internet access after a few days. (all firmwares i have tested have stability problems: 1.20, 1.21 and 1.32b02)
In the first two firmwares the problems started with īDNS ALG relay message problemsī All clients donīt have access to DNS anymore. so no internet for all clients that get DNS settings via DHCP. (with or without DNS relay setting!!! yes really; I also tested this by setting fixed DNS settings in the client and switching of DNS relay, it can just not contact the DNS servers anymore)
1.32b02 has a new issue that pops up first after a few days: it shuts down wifi. After it has done that you can also not log in to the router anymore. It continously returns you to the log in page if you have logged in correctly.
Stablilty seems related to how much clients-data you ask the router to handle. So the harder the router has to work the sooner it decides to have itīs beauty sleep.
As there seem to be time issues with this router I think my next test will be to deactivate the schedule. I have a schedule that switches of wifi at night. If time is messed up it might explain that wifi is cut off during the day but doesnīt explain why I can not log in to the router from that point on until I reset it.
I hope a TS has enough information to start testing with this info. I am happy to share log or setting files if that would help (to a real TS, not to offend you Clancy, just have to say it I guess else you get pissed again ;))
-
Just turned of the schedule and noticed something really odd:
On the status page of the router is says:
Mon Jun 22 2009 21:50:54 GMT+0200 (Central Europe Daylight Time)
i just updated the firmware about a week ago and copied the time from my computer via the button on the time page.
Just did it again and saved it, the date is now correct.
it seems within a week it added an hour and lost a few days???
-
it seems within a week it added an hour and lost a few days???
I think it shows you the time from the last change when it wrote the settings into its non volatile memory.
Did you change settings and save them at that time ?
-
After upgrading the default time is January first a few years ago.
As I also discovered an NTP issue the only thing I did was copy the time from my computer (via the button on the time page) just after upgrading about a week ago
So I did change and save it else the time would not be that recent?
Exactly an hour difference suggests time offset bugs in the software.
The few days difference I can not explain at all
b.t.w. the time is shown in the overall status page and the time page, both where showing the same time and date, which was wrong
-
Just (23:40) checked the time of my router, set it 2.5 hours ago.
On the status page it is already running 35 minutes behind: 23:05 :o
but when going to the time page it looks ok again?!
Reverting back to the time page shows the correct time again??
The log doesn't show much except:
[INFO] Sun Jun 28 23:16:44 2009 Administrator logout
i do not see an Admin login message?
So I logged in after 23:16, it showed me 23:05 when in fact, at the time i looked it was 23:40
Any explanation for that?
-
So... I have disabled all NTP time options, Set the wireless to 802.11g only using Channel 11 (no auto scan), disabled DNS Relay, disabled Shareport and disabled all wireless scheduling options (set to Always On) on my 1.32b02 A3 Router and now it's been running steady for 3 days and 15 hours without a restart!!!! WoooHoooo!
First of all, I wish I knew what combination of the above actually fixed the restart problem.
Second, if I knew that I had to disable most advanced features of this router, I wish I just bought a $45 Wireless-G router with no features instead.
At least then I wouldn't be offering banana's to everyone at low low prices!
jB
-
Hey Johnny. I have the same setup as you except I do have N+G enabled and it hasn't gone down in almost a week, Try enabling that and change nothing else.
-
Johnnyboy, if you have it running for a while without issues. Add 1 thing back and see if it stays running. Wait a few days to a week then add 1 more and see what happens. You may be able to isolate which combination of things is causing the reboots.
-
Yeha, that's what I'm doing now. I just added the 802.11n+g option on channel 11. I still think it's got something to do with the NTP, but *sometimes* the NTP serivice runs without a hitch, which is why I am looking into combination factors now.
I think if it can go 48 hours without restarting itself, I will consider it stable.
jB
-
I have also seen NTP issues as said in a previous thread. I can safely say that also pool.ntp.org gives problems so indeed it does not seem to matter which server you use.
You should not use just pool.ntp.org as that is all of the servers so your router keeps checking them till it finds the one for your time zone. This can result in your IP being banned by the servers.
From the pool site:
As pool.ntp.org will assign you timeservers from all over the world, time quality will not be ideal. You get a bit better result if you use the continental zones (http://www.pool.ntp.org/zone/@) (For example europe (http://www.pool.ntp.org/zone/europe), north-america (http://www.pool.ntp.org/zone/north-america), oceania (http://www.pool.ntp.org/zone/oceania) or asia (http://www.pool.ntp.org/zone/asia).pool.ntp.org), and even better time if you use the country zone (like ch.pool.ntp.org in Switzerland) - for all these zones, you can again use the 0, 1 or 2 prefixes, like 0.ch.pool.ntp.org. Note, however, that the country zone might not exist for your country, or might contain only one or two timeservers. If you know timeservers that are really close to you (measured by network distance, with traceroute or ping), time probably will be even better.
-
I disabled the schedule option as well and it is now running for 3 days which is a record for the thing!
My feeling is also that the time settings are the biggest suspect.
NTP, schedules, time zones, etc.
Followed by Wifi...
It has shown for me (can any one else check this?) saving the config with NTP settings and reverting back from factory default to those settings goes wrong. Without NTP settings in the config all seems ok....
-
I finally have some proof that there is a problem with this router regarding time.
As posted in the NTP thread I set the time of my router a few days ago.
I switched many features of and it has been running for 3 days which for the thing is a miracle.
Attached a screenshot just taken.
it says July 2nd just after nine pm. So in about 4-5 days it runs almost two days behind.
Let's just say an unacceptable margin....
Can somebody please do a check at their side for this. The settings are shown in the displayed jpg.
LOL even the website doesn't work correctly... I will upload on the web and insert picture in a minute
(http://i174.photobucket.com/albums/w112/partach/router4.jpg)
for those that don't believe I pressed the copy time from computer button before capturing the image
(http://i174.photobucket.com/albums/w112/partach/router2.jpg)
Thinking my English was bad..... look at the alive time spelling....
-
Just activated NTP to correct my time and guess what!
INSTANT DEATH. >:(
The 655 got into a continious reboot cycle. No control possible. Only option: factory reset on the back...
This proves it beyond doubt for me that there is a serious problem with this thing.
(http://i174.photobucket.com/albums/w112/partach/router5.jpg)
and of-course the log is gone !"!#$!"$
-
and of-course the log is gone !"!#$!"$
Thats why you should use something like THIS (http://www.softpedia.com/get/Internet/Servers/Server-Tools/Router-Syslog.shtml)
-
1.31 did show these time-symptoms (lagging behind), 1.32 does not anymore.
-
did you see the picture with the version displayed? it clearly says 1.32.....
Hence 1.32 does....
-
Well, just to be sure, for now do not use pool.ntp.org. There is too much uncertainty about that NTP service with the dlink routers.
For now, just stick with the default D-Link NTP service, because honestly, there is no frickin' difference. That should prevent your router restarts.
jB
-
For now, just stick with the default D-Link NTP service, because honestly, there is no frickin' difference. That should prevent your router restarts.
jB
That's the best solution, but some peeps seem to get off on finding things that make a router crash (which really isn't that hard, with any brand) :D
-
Demonized would you just quit the innuendo, having this router is bad enough.
ntp=ntp=ntp
using time servers from ntp.pool.org for decades, always worked fine. Other client in my network uses it for two years now, never crashed once.
just trying to use a localized freakin server, too bad dlink makes up their own ntp protocol (which I doubt).
as shown before the router handles time incorrectly, let's just let them make a fix for that as demonized said it was in 131 and I have shown it is in 132 as well (never used 131 myself)
-
Demonized would you just quit the innuendo, having this router is bad enough.
ntp=ntp=ntp
using time servers from ntp.pool.org for decades, always worked fine. Other client in my network uses it for two years now, never crashed once.
just trying to use a localized freakin server, too bad dlink makes up their own ntp protocol (which I doubt).
as shown before the router handles time incorrectly, let's just let them make a fix for that as demonized said it was in 131 and I have shown it is in 132 as well (never used 131 myself)
That's really too bad, my DIR655 works really great. So no innuendo, you having the problem does not automatically make it a bad router, it takes a bit more to achieve that status. And I've never seen a badly designed machine perform very well as an exception. Have you?
-
That's really too bad, my DIR655 works really great. So no innuendo, you having the problem does not automatically make it a bad router, it takes a bit more to achieve that status. And I've never seen a badly designed machine perform very well as an exception. Have you?
What does that even mean?
Yes some badly designed machines do in fact perform really well. Certain remotes for example seem not to be designed as intuitively user friendly but do exceptional jobs once the user adjusts.
Old Corvair cars, and 80's Subarus with the faulty body metal.
And the Bumble bee.
I would say your perfect router is the exception NOT the rule. That is further confirmed by the poll on router problems.
-
What does that even mean?
Yes some badly designed machines do in fact perform really well. Certain remotes for example seem not to be designed as intuitively user friendly but do exceptional jobs once the user adjusts.
Old Corvair cars, and 80's Subarus with the faulty body metal.
And the Bumble bee.
I would say your perfect router is the exception NOT the rule. That is further confirmed by the poll on router problems.
Yeah, it's a very reliable poll; 10 people with issues out of 100.000+ sold devices...wow. Statistically that's what they call 'marginal'.
Because you're device is troublesome it really does not prove all 100.000 devices are defects or the design is flawed. That's a very self centered conclusion and calling a well functioning device the exception is based on very thin air.
Just figure out the math: how many people visiting the forum have claimed they have an issue and how many just come here to say Hi and have no issues? 99% are issue people, you can count them if you like. So a poll will reflect this number and therefore the poll result will say 99% has an issue. It's a false conclusion to take that result and qualify the router as 'design flawed'. Emperically this falls in the category "the earth is flat".
-
... Emperically this falls in the category "the earth is flat".
It's not ?
:o
-
We can talk about it if you want... ;D
-
Don't know why you all "fight" aggressively for a statement that is only to be seen relatively.
There are DIR-655 in special environment and/or use which don't work propperly. This is fact.
But there are also 655 working flawlessly as mine. It only needed one restart since the last firmware because I exchanged the WAN connection from cable modem to DSL modem. (To test 655 and 855 in both scenarios).
The number of working and not working routers doesn't matter. If there is a problem, it has to be solved.
In this case , I think it is D-Link's work to release a firmware which is stable in all environments.
That's all !
btw.
The "flat earth" was a joke and I read the answer as a joke too.
Not all is meant seriously...
-
CENSORED
Looking at your vocabulary you are not really a civilized person. You're seeing things that aren't there and seem to have a problem understanding emoticons.
Thanks lizzi555 for understanding, we were indeed just having some fun.
I'm sorry I'm ruining your Dlink bashing though. :P
PS: your 'smite' score is getting close to mine, but without the number 'applauds'. Howcome...?
*** Modified by Fatman because I didn't see a point in keeping that quote around. Truthfully I am not even sure why I am leaving the response to it, other than not having the same level of reason to remove it.
-
Lets keep things friendly guys, no blows below tha waist. We tend to tolerate heated discussions as long as they don't turn personal or make one unreasonable, both happened here, so we won't be hearing from GullyFoyle for 3 days. At the same time I want to remind everyone that non-constructive threads get locked eventually and that if you all just want to duke it out, head outside please.