• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

This one seems to be a stumper for me. Need suggestions to resolve disconnects

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

notarat

Member
Joined
Nov 11, 2010
Some time back I posted about a connection problem I was experiencing with Path of Exile(POE)

I ran WinMTR when these disconnects were occurring. WinMTR does a fancy traceroute so you can see all the hops between your computer and it logs connection statistics for analysis. I found that if I had 100 disconnects in a day, 75% of them occurred on the servers where POE is hosted...but a full 25% occurred on my end.

EXAMPLE:
Conx1 - comp to router (25% of all disconnects occur here)
Conx2 - Router to Cable Modem
Conx3 - Cable Modem to Comcast
Conx4 through 10 (let's use 10 as an example of the number of hops) different servers/routes between comcast and the hosting company for the POE Servers
Conx 11 - hosting server of POE (75% of all disconnects occur here)

The "symptom" I see is the network icon in the system tray displays the red X through the Network connection, like when you unplug your network cable. It's up, then it's down, then it's up, then it's down, repeat, repeat, repeat for up to 60 minutes at a time. Never at a certain time. Never on a certain day of the week, never on a certain date each month
Event viewer only displays a generic message stating the connection was lost/restored. I've researched the event IDs and have gotten nowhere. The pages at Microsoft's site for these event IDs are very generic to the point of being useless.

Note: It is not a cable-related issue. I have swapped out my network cables several times with cables made by different mfgrs as well as using my own self-built CAT6 and CAT5e cables

My initial thought was that my problem was due to a bad/dying router...I believed that because the dropped connections on my end occurred before the traceroute reached the cable modem.

I replaced the router, and am still having the same problem with the same frequency of occurrences.

So...I thought to myself, "Maybe the issue is the onboard Intel NIC.", so I disabled it in BIOS and installed a different Intel Gigabit NIC Card based on an older Intel chip which used a completely different driver package. Same problem persists...


I replaced the Intel NIC Card with a spare Rosewill I had lying around. Same problem persists...

At this point I'm starting to believe it's something in the network setup of the OS on my computer itself so I'm looking for tips on what options I can tweak or disable to narrow down where/why this is happening.

Suggestions?
 
Ill try my hand at this, I know back when I worked at the call center there was a hard limit on the number of active IP sessions any given customer could have before a lockout would occur but what you're describing sounds LAN side. This might be a stupid question but you replaced the router with what? I had issues similar to what you describe and isolated it to a specific Broadcom SOC that was notorious for overheating and had to put a special edition netgear router to get one with a Qualcomm chipset and my problems went away.
 
Ill try my hand at this, I know back when I worked at the call center there was a hard limit on the number of active IP sessions any given customer could have before a lockout would occur but what you're describing sounds LAN side. This might be a stupid question but you replaced the router with what? I had issues similar to what you describe and isolated it to a specific Broadcom SOC that was notorious for overheating and had to put a special edition netgear router to get one with a Qualcomm chipset and my problems went away.

I've used the following 3 routers when trying to eliminate/ameliorate the disconnect issues:

Asus Black Knight RT66
Linksys WRT1900ACS
TPLink Archer C7

All have their wireless connectivity disabled.

Only 1 computer is "on" and on the network, and it is using a wired connection.

I believe each one uses a different chip, though I may be mistaken. (If they are using the same, it could explain why the problem persists even after changing NIC Cards and/or cables.)
 
I've used the following 3 routers when trying to eliminate/ameliorate the disconnect issues:

Asus Black Knight RT66
Linksys WRT1900ACS
TPLink Archer C7

All have their wireless connectivity disabled.

Only 1 computer is "on" and on the network, and it is using a wired connection.

I believe each one uses a different chip, though I may be mistaken. (If they are using the same, it could explain why the problem persists even after changing NIC Cards and/or cables.)


TP-LINK Archer C7 v2.x - WikiDevi
https://wikidevi.com/wiki/TP-LINK_Archer_C7_v2.x
Aug 10, 2016 - WI1 chip1: Qualcomm Atheros QCA9558 WI1 802dot11 protocols: bgn. WI1 MIMO config: 3x3:3. WI1 antenna connector: none. WI2 module: ...

ASUS
WI1 chip1: Broadcom BCM4331
WI1 802dot11 protocols: bgn
WI1 MIMO config: 3x3:3
WI1 antenna connector: U.FL, RP-SMA
WI2 chip1: Broadcom BCM4331
WI2 802dot11 protocols: an
WI2 MIMO config: 3x3:3
WI2 antenna connector: U.FL, RP-SMA


Linksys
WI1 module: unknown
WI1 module IF: Mini PCIe (dual)
WI1 chip1: Marvell 88W8864
WI1 802dot11 protocols: an+ac
WI1 MIMO config: 4x4:3
WI1 antenna connector: RP-SMA
WI2 module: unknown
WI2 module IF: Mini PCIe (dual)
WI2 chip1: Marvell 88W8864
WI2 802dot11 protocols: bgn
WI2 MIMO config: 4x4:3
WI2 antenna connector: RP-SMA



-----

Three different router chipsets... what modem are you using? I had a lengthy discussion with Comcast and apparently they're heavily pushing 3.1 for more pair bonds to multiplex on my guess is that either that is getting overwhelmed or you might want to look at carrier class routers like Ubiquiti or Cisco Aironets
 
Back