[Stigma]

Hey all, I need some help troubleshooting.

I'm getting a problem with my router (see title) where it seems to "jam" intermittently. It runs fine until it suddently starts to respond REALLY slowly, and pinging the router (even via cable) gives me up to 4000ms responstime, or sometimes not a response at all. Actual throughput drops to nearly nothing. This does not seem related to heavy network load either. The problem can eventually go away on its own. Otherwise a router reboot clears it (but it can sometimes return within minutes... or hours). I really need to get this fixed as it completely kills my online gaming

A few things happened recently that might be related:

- I recently upgraded firmware from the very old 15778 up to 23320, and then finally to 25179 to try to fix this issue (25179 was installed with a full hard-reset before and after flash, and manual re-setup which was basically only my PPPoE settings + wifi settings)

- I set up my wifi for 300Mbps support (its automatic in 25179 apparently)

- A few family members are over for chrismas so there are a handful more wifi clients active. Its still only about 8-9 at a time, and not massive traffic by any means.

The status page in DD-WRT looks fine. I have plenty of hardware resources, there aren't a gazillion connections being used ect. and the only one in the household that does any kind of filesharing or heavy downloading is me.

Anecdotally the problem seems to occur more when there are many wifi devices online (but not really all that much traffic, most of them are idle phones and such). I can't remember having this issue in the past (before the first firmware upgrade).

I haven't yet tried going all the way back to my old firmware + old setting, but that's essentially just a shot in the dark on my part so I wanted to ask how I can troubleshoot this properly so I can hopefully save myself days of testing each possible permutation. Also im not all that keen on running years old firmware if I don't absolutely have to.

Let me know if I can provide any more details that might help.

-Stigma

[KrypteX]

I'll try to help you out with this, since I have the D2 version and some experience with this router. All D versions have the same firmware.

I've used the router for ~1 year with mostly 2 firmwares: 23320 for 10 months, with the 300 Mb/s fix, and 25179 for >1 month, both with regular PPPoE config. I've also used 25143 for only 1-2 days, before I switched to 25179.

On 23320, the router run without any hiccups for very long times (several months at a time), but on 25179, even though it's faster, it locked up after 31 days (I've posted a thread about this). I've since changed the router with a Netgear.

Things that could be the culprit:
TX Power --- I've always set it to 71 mW, even though it's @ 100 mW by default in builds starting with 23503. I think that's too high, since the router manual says it's designed for approx 40 mW. It may be that you've damaged your router by running it @ 100 mW. BTW, you couldn't have run that high for a long time, because 25179 is out for ~ 1-2 months and you said you didn't use anything between 23320 and 25179 for an extended period.

Another problem might be the WiFi channel. I recommend setting a fixed channel, not Auto. I've run this router for the last year or so on Channel 11, never had serious problems with it.

To troubleshoot, I'd flash back 23320 and test for a while @ 71 mW and fixed Channel (use 1 or 6 or 11). These are my settings that I usually do after a factory reset:

1. Setup -> Basic Setup -> WAN Connection: PPPoE, Username, Password
-> Router Name: DIR-615
-> Time Zone: Europe/City. Save.

2. Wireless -> Basic settings -> 2.4G Phys ra0: AP, Mixed, HT40, Ch 11, DIR-615. Save.

3. Wireless -> Wireless Security: WPA2 Personal, AES, Password. Save.

4. Wireless -> Advanced Settings -> TX Power: 71 mW. Save.

5. ONLY for builds < 25143: Administration -> Commands -> add these lines:

sleep 20
iwpriv ra0 set HtTxStream=2
sleep 5
iwpriv ra0 set HtRxStream=2

Save startup.

6. Administration -> Management -> Apply Settings.
Wait 1-2 minutes for changes to take effect. Reboot Router.

[Stigma]

I appreciate any help I can get Kryptex

I haven't run high TX power. It's been at 72 (i think, aproximately) as a default of the older firmwares. In the new ones (which I only started messing with a few days before christmas) I've tuned it manually and found I got the best result around 92 anyway, so I left it there. Don't think that's likely to burn out the router or anything - and in any case I don't think a weakened transmitter would be a likely culprit for making the whole router jam up like that even for the cabled traffic (and work fine for the most part after all)

wifi channel is set manually to 6 (40-wide, upper). The other networks in the area are pretty faint and I barely have any overlap with any other signal.

You really think this problem could have to do with the wifi at all though? I mean, if that's the case then I can always just try to change my settings down to a very basic level - low TX power, narrow manual channel ect. and see if that changes anything. Honestly I just don't know if problems related to wifi have the potential to gum up the whole rotuer like this, but if you think so then I can give that a shot.

Part of the problem in troubleshooting is that the problem can not present itself for long periods of time. It seems to happen the most when there are a bit more clients on than usual - but I have no idea if this is due to extra strain locking up the router somehow, or perhaps just one particular client having some kind of bad reaction with my router (I think all of the extra clients that are here for the holidays are various apple devices). I think I will try pinglogging overnight just to see if the problem manifests itself during the night when there is little to no activity...

Oh, and I had the problem on 23320 too (or rather that's the first firmware I noticed the problem on), so I don't see much point in flashing back to that one.

-Stigma

[KrypteX]

It also might be due to an increased number of IP connections, when you have ~10 clients. If you check Status -> Router, at the bottom of the page you'll see the max allowed total number of ports/connections. Default is 4096. Below that, you'll see the actual number in use.

You can equally check the number of connections for each client in Status -> LAN -> Active Clients -> Conn. Count.
I've noticed a relatively large number for clients running torrent and facebook at the same time -- usually around 1300 per client. That can quickly add up to 4096 and more.

You could try setting Administration -> Management -> IP Filter Settings -> Maximum Ports to something higher, let's say 10000 or 15000. Interestingly, on both of my routers, the default setting is 4096, even though in paranthesis, it says: Default: 32768. Hmmm....

[Stigma]

Its happening right now and im at less than 7% of max connections (under 300). it also happened a couple of times through the middle of the night too (no users, only idle devices).

-Stigma

[KrypteX]

Try 25143. If issue persists, go back to original D-Link firmware from Emergency Room and try with that for a while, to see if router is damaged or if it's just DD-WRT that gives you troubles.

[Stigma]

I'll try returning to the exact config that I was using before first (I know the build + i made a config backup right before I started messing with it). I'll see how that makes the router behave.

If that still has the issue I'll maybe check out the original firmware like you suggest. Obviously if the problem happens there too then there is likely something wrong with the rotuer itself - I just kind of doubt it given that this all started when I made firmware changes.

I'll keep you posted, and thanks for the advice

-Stigma

[Stigma]

ok im now back on the "original" 15778 firmware with my old tried and true settings. Gonna run this for a day or two - or until I see the same problem manifest itself again (I'm logging 24/7 to check for it).

We'll see what happens here. I'm not really content with running firmware from 2011 with poor support for 300mbps, but at least if I can confirm this is stable then I have a "safe fallback" solution while I experiment with other builds - and if this is stable I guess it would be fair to conclude that there isn't anything wrong with the router itself at least.

EDIT: I immediately notice going back from 25179 that the UI is now snappy and responsive again. It was only slow on 25179 out of the ones I've tested so far though (sometimes 3-5 seconds to load a page).

EDIT2: Exactly what version of the DIR-615 did you have KrypteX? I'm running a D1, do we have identical units, or just both D-series (ie. same firmware)?

EDIT3: It seems likely that Im just going to have to try out a bunch of the later firmwares to see if any of them work correctly for me (assuming that 15778 continues to be stable for now). You seem pretty active on the forums KrypteX, do you have a list of the known problematic ones that have the potential for bricking?

Will update when stuff happens...
-Stigma

[KrypteX]

It's a good idea to test that old 15778 build first, to be sure that there's no damage to the router in the first place (hopefully everything will turn out fine).

I have the DIR-615 D2 version myself, which I guess it's more or less the same hardware as D1. According to this page
http://www.dd-wrt.com/wiki/index.php/Supported_Devices
they look the same to me. Anyway, the firmware for all D versions (D1-D4) is exactly the same.

Regarding 25179, I have never seen Webgui slowdowns with it, it's snappy for me, it's not slower than 23320, for example. Try using the latest Firefox and clear its cache (Options-Advanced-Network-Cached Web Content-Clear Now) and restart Firefox. This usually does the trick.

On the other hand, it looks like BrainSlayer has found the bug which bricked all firmwares after 25179 for this router. We've been discussing about this today and I'm testing thoroughly right now. See more details in this ticket (you'll even see a new experimental build right there - be careful though with flashing it):
http://svn.dd-wrt.com/ticket/3707
Let's hope the next official build (after 25697) will fix most of the remaining issues with DIR-615 D versions.

[KrypteX]

I've just finished testing the test build provided by BrainSlayer for DIR-615 D versions, which solves the softbrick on reboot. It's build 25737 from 12/28/2014 and can be downloaded here:
http://svn.dd-wrt.com/ticket/3707#comment:33
I've tested it in AP, Client and Repeater modes, works wonderfully !

[Stigma]

Well... even the old firmware&config still gets the problem. How frustrating...

The most I test the more I am convinced it has something to do with the wifi (or the wifi clients). It definitely tend to happen when there are many (7-10) clients connected, and it seems to be completely unrelated to how much load is on the network - as it can occur while everyone is just idling.

I'm not sure what this means... it could mean that my DIR-615 for som reason just breaks down completely and can't handle 8-9-10 active clients, or it could maybe be ONE of the clients that is acting up and somehow affecting the router... or I suppose my DIR-615 could just be dying, but I'm not willing to default to that conclusion quite yet :/

EDIT: I'm currently in the process of trying to weed out each client as a possible troublemaker by shutting them physically off while watching my pings while the problem is active. I guess if I find one that consistently removes the issue every time I turn it off then I might have my culprit. That's assuming that there actually exists a "troublemaker" client.

-Stigma

[KrypteX]

Flash this new build first, it works fine for me for the last 24h:
http://svn.dd-wrt.com/ticket/3707#comment:33

[Stigma]

Ok I am now on build 25737. Monitoring ... let's see how this one turns out.

EDIT: Well, so far so good. 8-9 clients with sporadic traffic. Will try to stress the network in a little while with as many units as I can to see if I can provoke anything to happen.

-Stigma

[Stigma]

Ok, I've been running a couple of days now on the new build, and this seems to have fixed it. I still have some very minor hickups detected (50-300ms for a second or two) but its nothing remotely like what I had before and might be entirely unrelated to the previous issue.

Thanks a ton for the assistance. It's going to grate on me to now know exactly what was causing the issue - but having a network that actually works certainly helps a lot

-Stigma

[KrypteX]

It's great to hear good news from time to time ! Happy New Year !!!