why this caused whole network hang? [7:60090]
Hi.. I have one of my 2900 switch connected to my backbone switch(5500). A few day ago, I found that the switch to switch connection has duplex mismatch error. The 2900 sw interface was set to 100 full dpx. But the 5505 sw interface was set to auto neg, but they fail to negotiate properly. Therefore, I tried to set the 5505 sw interface to 100 full dx. But immediately after I set, the whole company network hang. All users lost connection for more than 10 minutes. I really can't understand why?? The 2900 sw has only one connection to the 5505 sw and both interface set to portfast (faststart) disable. I think it is related to spanning-tree but why spanning-tree loop will happen in this case. Can one explain to me. Thanks a lot _ Protect your PC - get McAfee.com VirusScan Online http://clinic.mcafee.com/clinic/ibuy/campaign.asp?cid=3963 Message Posted at: http://www.groupstudy.com/form/read.php?f=7&i=60090&t=60090 -- FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
RE: why this caused whole network hang? [7:60090]
A duplex mismatch problem somewhere in the network can indeed cause problems for STP. It's an example of the infamous unidirectional link problem, which is also known as the one-way connectivity failure. The side that is using full-duplex can send OK because it ignores the fact that the other side is sending at the same time. The side that is using half-duplex gets excessive collisions and gets so busy handling those that it can't send or receive frames reliably. So it doesn't hear BPDUs from the other side. This can result in it setting one of its ports into fowarding mode when it souldn't. Cisco has a feature to avoid unidirectional link problems. Try looking up their Unidirectional Link Detection protocol, although the real fix in this situation isn't to use the protocol, but, rather to fix the duplex problem, as you know. Priscilla Kenny Smith wrote: > > Hi.. I have one of my 2900 switch connected to my backbone > switch(5500). A > few day ago, I found that the switch to switch connection has > duplex > mismatch error. The 2900 sw interface was set to 100 full dpx. > But the 5505 > sw interface was set to auto neg, but they fail to negotiate > properly. > Therefore, I tried to set the 5505 sw interface to 100 full > dx. But > immediately after I set, the whole company network hang. All > users lost > connection for more than 10 minutes. > > I really can't understand why?? The 2900 sw has only one > connection to the > 5505 sw and both interface set to portfast (faststart) > disable. I think it > is related to spanning-tree but why spanning-tree loop will > happen in this > case. Can one explain to me. > > Thanks a lot > > > _ > Protect your PC - get McAfee.com VirusScan Online > http://clinic.mcafee.com/clinic/ibuy/campaign.asp?cid=3963 > > Message Posted at: http://www.groupstudy.com/form/read.php?f=7&i=60130&t=60090 -- FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
RE: why this caused whole network hang? [7:60090]
Hi.. Priscilla and Dear all, Thanks for your explaination and sorry for my ignorance. But I don't understand one part about "it doesn't hear BPDUs from the other side. This can result in it setting one of its ports into fowarding mode when it souldn't." But my switch has only one connection to my backbone switch. So I don't understand how it will creat STP loop? >From: "Priscilla Oppenheimer" >Reply-To: "Priscilla Oppenheimer" >To: [EMAIL PROTECTED] >Subject: RE: why this caused whole network hang? [7:60090] >Date: Thu, 2 Jan 2003 21:44:06 GMT > >A duplex mismatch problem somewhere in the network can indeed cause >problems >for STP. It's an example of the infamous unidirectional link problem, which >is also known as the one-way connectivity failure. > >The side that is using full-duplex can send OK because it ignores the fact >that the other side is sending at the same time. The side that is using >half-duplex gets excessive collisions and gets so busy handling those that >it can't send or receive frames reliably. So it doesn't hear BPDUs from the >other side. This can result in it setting one of its ports into fowarding >mode when it souldn't. > >Cisco has a feature to avoid unidirectional link problems. Try looking up >their Unidirectional Link Detection protocol, although the real fix in this >situation isn't to use the protocol, but, rather to fix the duplex problem, >as you know. > >Priscilla > > >Kenny Smith wrote: > > > > Hi.. I have one of my 2900 switch connected to my backbone > > switch(5500). A > > few day ago, I found that the switch to switch connection has > > duplex > > mismatch error. The 2900 sw interface was set to 100 full dpx. > > But the 5505 > > sw interface was set to auto neg, but they fail to negotiate > > properly. > > Therefore, I tried to set the 5505 sw interface to 100 full > > dx. But > > immediately after I set, the whole company network hang. All > > users lost > > connection for more than 10 minutes. > > > > I really can't understand why?? The 2900 sw has only one > > connection to the > > 5505 sw and both interface set to portfast (faststart) > > disable. I think it > > is related to spanning-tree but why spanning-tree loop will > > happen in this > > case. Can one explain to me. > > > > Thanks a lot > > > > > > _ > > Protect your PC - get McAfee.com VirusScan Online > > http://clinic.mcafee.com/clinic/ibuy/campaign.asp?cid=3963 _ MSN 8 with e-mail virus protection service: 2 months FREE* http://join.msn.com/?page=features/virus Message Posted at: http://www.groupstudy.com/form/read.php?f=7&i=60170&t=60090 -- FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
RE: why this caused whole network hang? [7:60090]
Kenny Smith wrote: > > Hi.. Priscilla and Dear all, > > Thanks for your explaination and sorry for my ignorance. But I > don't > understand one part about "it doesn't hear BPDUs from the other > side. This > can result in it setting one of its ports into fowarding mode > when it > souldn't." But my switch has only one connection to my > backbone switch. So > I don't understand how it will creat STP loop? Well maybe it wasn't a loop. But a duplex mismatch problem will cause BPDUs to flow inconsistently which could cause STP to reconverge incessantly or other problems. I wasn't there and I don't know your topology, so I can't say for sure what happened. Priscilla > > > >From: "Priscilla Oppenheimer" > >Reply-To: "Priscilla Oppenheimer" > >To: [EMAIL PROTECTED] > >Subject: RE: why this caused whole network hang? [7:60090] > >Date: Thu, 2 Jan 2003 21:44:06 GMT > > > >A duplex mismatch problem somewhere in the network can indeed > cause > >problems > >for STP. It's an example of the infamous unidirectional link > problem, which > >is also known as the one-way connectivity failure. > > > >The side that is using full-duplex can send OK because it > ignores the fact > >that the other side is sending at the same time. The side that > is using > >half-duplex gets excessive collisions and gets so busy > handling those that > >it can't send or receive frames reliably. So it doesn't hear > BPDUs from the > >other side. This can result in it setting one of its ports > into fowarding > >mode when it souldn't. > > > >Cisco has a feature to avoid unidirectional link problems. Try > looking up > >their Unidirectional Link Detection protocol, although the > real fix in this > >situation isn't to use the protocol, but, rather to fix the > duplex problem, > >as you know. > > > >Priscilla > > > > > >Kenny Smith wrote: > > > > > > Hi.. I have one of my 2900 switch connected to my backbone > > > switch(5500). A > > > few day ago, I found that the switch to switch connection > has > > > duplex > > > mismatch error. The 2900 sw interface was set to 100 full > dpx. > > > But the 5505 > > > sw interface was set to auto neg, but they fail to negotiate > > > properly. > > > Therefore, I tried to set the 5505 sw interface to 100 full > > > dx. But > > > immediately after I set, the whole company network hang. All > > > users lost > > > connection for more than 10 minutes. > > > > > > I really can't understand why?? The 2900 sw has only one > > > connection to the > > > 5505 sw and both interface set to portfast (faststart) > > > disable. I think it > > > is related to spanning-tree but why spanning-tree loop will > > > happen in this > > > case. Can one explain to me. > > > > > > Thanks a lot > > > > > > > > > > _ > > > Protect your PC - get McAfee.com VirusScan Online > > > http://clinic.mcafee.com/clinic/ibuy/campaign.asp?cid=3963 > _ > MSN 8 with e-mail virus protection service: 2 months FREE* > http://join.msn.com/?page=features/virus > > Message Posted at: http://www.groupstudy.com/form/read.php?f=7&i=60175&t=60090 -- FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
RE: why this caused whole network hang? [7:60090]
Thanks Priscilla, I understand what you mean now. That's why those ports which directly connected to PCs need to configured as spanning-tree portfast to minimize the STP state change. >From: "Priscilla Oppenheimer" >Reply-To: "Priscilla Oppenheimer" >To: [EMAIL PROTECTED] >Subject: RE: why this caused whole network hang? [7:60090] >Date: Fri, 3 Jan 2003 04:01:41 GMT > >Kenny Smith wrote: > > > > Hi.. Priscilla and Dear all, > > > > Thanks for your explaination and sorry for my ignorance. But I > > don't > > understand one part about "it doesn't hear BPDUs from the other > > side. This > > can result in it setting one of its ports into fowarding mode > > when it > > souldn't." But my switch has only one connection to my > > backbone switch. So > > I don't understand how it will creat STP loop? > >Well maybe it wasn't a loop. But a duplex mismatch problem will cause BPDUs >to flow inconsistently which could cause STP to reconverge incessantly or >other problems. I wasn't there and I don't know your topology, so I can't >say for sure what happened. > >Priscilla > > > > > > >From: "Priscilla Oppenheimer" > > >Reply-To: "Priscilla Oppenheimer" > > >To: [EMAIL PROTECTED] > > >Subject: RE: why this caused whole network hang? [7:60090] > > >Date: Thu, 2 Jan 2003 21:44:06 GMT > > > > > >A duplex mismatch problem somewhere in the network can indeed > > cause > > >problems > > >for STP. It's an example of the infamous unidirectional link > > problem, which > > >is also known as the one-way connectivity failure. > > > > > >The side that is using full-duplex can send OK because it > > ignores the fact > > >that the other side is sending at the same time. The side that > > is using > > >half-duplex gets excessive collisions and gets so busy > > handling those that > > >it can't send or receive frames reliably. So it doesn't hear > > BPDUs from the > > >other side. This can result in it setting one of its ports > > into fowarding > > >mode when it souldn't. > > > > > >Cisco has a feature to avoid unidirectional link problems. Try > > looking up > > >their Unidirectional Link Detection protocol, although the > > real fix in this > > >situation isn't to use the protocol, but, rather to fix the > > duplex problem, > > >as you know. > > > > > >Priscilla > > > > > > > > >Kenny Smith wrote: > > > > > > > > Hi.. I have one of my 2900 switch connected to my backbone > > > > switch(5500). A > > > > few day ago, I found that the switch to switch connection > > has > > > > duplex > > > > mismatch error. The 2900 sw interface was set to 100 full > > dpx. > > > > But the 5505 > > > > sw interface was set to auto neg, but they fail to negotiate > > > > properly. > > > > Therefore, I tried to set the 5505 sw interface to 100 full > > > > dx. But > > > > immediately after I set, the whole company network hang. All > > > > users lost > > > > connection for more than 10 minutes. > > > > > > > > I really can't understand why?? The 2900 sw has only one > > > > connection to the > > > > 5505 sw and both interface set to portfast (faststart) > > > > disable. I think it > > > > is related to spanning-tree but why spanning-tree loop will > > > > happen in this > > > > case. Can one explain to me. > > > > > > > > Thanks a lot > > > > > > > > > > > > > > _ > > > > Protect your PC - get McAfee.com VirusScan Online > > > > http://clinic.mcafee.com/clinic/ibuy/campaign.asp?cid=3963 > > _ > > MSN 8 with e-mail virus protection service: 2 months FREE* > > http://join.msn.com/?page=features/virus _ STOP MORE SPAM with the new MSN 8 and get 2 months FREE* http://join.msn.com/?page=features/junkmail Message Posted at: http://www.groupstudy.com/form/read.php?f=7&i=60188&t=60090 -- FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
RE: why this caused whole network hang? [7:60090]
Portfast doesn't turn off STP and state changes can still happen. It just puts the port into forwarding mode right away. The port still listens and reacts to BPDUs. Don't be tempted to disable STP. It leaves the network too vulnerable to loops. Even if you have a perfectly-shaped physical tree network already, a user or newbie technician could install a switch in such a way that your tree gets tangled. (I don't think you are tempted to turn it off, but I just thought I would add that recommendation just FYI.) Portfast is the right config to have on your access ports. I'm not sure you really want to have it on the switch-to-switch port, though. There are other features to speed up spanning tree resolution for switch-to-switch ports, such as Uplink Fast and Backbone Fast. Those may not be supported on the low-end switches though. What fixed the problem or did it just stop happening after 10 minutes? I don't think you'll ever know for sure what happened, unless you could reproduce the problem, which you wouldn't want to do! :-) Priscilla Kenny Smith wrote: > > Thanks Priscilla, I understand what you mean now. That's why > those ports > which directly connected to PCs need to configured as > spanning-tree portfast > to minimize the STP state change. > > > >From: "Priscilla Oppenheimer" > >Reply-To: "Priscilla Oppenheimer" > >To: [EMAIL PROTECTED] > >Subject: RE: why this caused whole network hang? [7:60090] > >Date: Fri, 3 Jan 2003 04:01:41 GMT > > > >Kenny Smith wrote: > > > > > > Hi.. Priscilla and Dear all, > > > > > > Thanks for your explaination and sorry for my ignorance. > But I > > > don't > > > understand one part about "it doesn't hear BPDUs from the > other > > > side. This > > > can result in it setting one of its ports into fowarding > mode > > > when it > > > souldn't." But my switch has only one connection to my > > > backbone switch. So > > > I don't understand how it will creat STP loop? > > > >Well maybe it wasn't a loop. But a duplex mismatch problem > will cause BPDUs > >to flow inconsistently which could cause STP to reconverge > incessantly or > >other problems. I wasn't there and I don't know your topology, > so I can't > >say for sure what happened. > > > >Priscilla > > > > > > > > > >From: "Priscilla Oppenheimer" > > > >Reply-To: "Priscilla Oppenheimer" > > > >To: [EMAIL PROTECTED] > > > >Subject: RE: why this caused whole network hang? [7:60090] > > > >Date: Thu, 2 Jan 2003 21:44:06 GMT > > > > > > > >A duplex mismatch problem somewhere in the network can > indeed > > > cause > > > >problems > > > >for STP. It's an example of the infamous unidirectional > link > > > problem, which > > > >is also known as the one-way connectivity failure. > > > > > > > >The side that is using full-duplex can send OK because it > > > ignores the fact > > > >that the other side is sending at the same time. The side > that > > > is using > > > >half-duplex gets excessive collisions and gets so busy > > > handling those that > > > >it can't send or receive frames reliably. So it doesn't > hear > > > BPDUs from the > > > >other side. This can result in it setting one of its ports > > > into fowarding > > > >mode when it souldn't. > > > > > > > >Cisco has a feature to avoid unidirectional link problems. > Try > > > looking up > > > >their Unidirectional Link Detection protocol, although the > > > real fix in this > > > >situation isn't to use the protocol, but, rather to fix the > > > duplex problem, > > > >as you know. > > > > > > > >Priscilla > > > > > > > > > > > >Kenny Smith wrote: > > > > > > > > > > Hi.. I have one of my 2900 switch connected to my > backbone > > > > > switch(5500). A > > > > > few day ago, I found that the switch to switch > connection > > > has > > > > > duplex > > > > > mismatch error. The 2900 sw interface was set to 100 > full > > > dpx. > > > > > But the 5505 > > > > > sw interface was set to auto neg, but they fail to > negotiate > > > > > properly. > > > > >
