- mice-discuss - lists.micemn.net

Re: EX4200 (1G Switch) Packet Loss -> 10G Participants

by Ryan Goldberg

No sfp slots available on the 4200. It can only run the two 10g... On Sep 29, 2016, at 10:50 AM, Matthew Beckwell <matthewb(a)AITECH.NET<mailto:matthewb@aitech.net>> wrote: Jeremy, Thanks for your efforts on this. Personally, i'm fine with next Tuesday afternoon. I think the general consensus was mid-day maintenance is fine, provided we send out something to MICE-ANNOUNCE (probably by the end of the day today) so people can shut down their sessions in advance if they prefer. I don't remember...did we decide if we're moving the 1G Optical folks on the existing 4200 expansion module (5 Nines, ClaimLynx, CDW/Berbee, Velocity Telephone) to the 4550? ~Matthew On Thu, Sep 29, 2016 at 8:35 AM, Jeremy Lumby <jlumby(a)mnvoip.com<mailto:jlumby@mnvoip.com>> wrote: As many of you know after Tuesday's meeting, I moved my connection over to the new 4200 that was donated by Compudyne. The topology is that the existing three switches are connected together in a ring with Juniper stacking cables, and then the new 4200 that I am testing has 2 10G ports lagged to the existing switches, one of which goes to the 4500, and the other goes to the 4550. When I initially connected, I was seeing loss to all three of the older switches. The loss was 0.4% to the 4500, and 0.2% to the 4200, and the 4550. I then shutdown the LAG port that ran between the 4550, and the new 4200. The new loss numbers over a 6 hour period came back as 0% to the 4500 and 4200, and 0.2% to the 4550. I then switched it so that only the port to the 4550 was active in the LAG group, and the loss changed to 0.3% to the 4500, 0.2% to the 4200, and 0% to the 4550. I believe this clearly points to the loss being in the stacking of the existing switches. My suggested next step would be to schedule a maintenance window where the existing members are moved from the old 4200 to the new one, and the LAG port from the new 4200 to the 4550 is left shutdown. This will enable the existing 4200 users to experience less loss, while we can see if removing the old 4200, and associated stacking cables has removed the loss from the existing stack. Because of my limited availability I am proposing doing the maintenance next Tuesday afternoon. Let me know if this works, or if someone else with better availability would rather do it. It should only briefly effect the 1G members, and not the 10G users. I plan on moving the route servers one at a time leaving plenty of time for BGP to re-establish prior to moving the second one. Jeremy Lumby Minnesota VoIP 9217 17th Ave S Suite 216 Bloomington, MN 55425 Main: 612-355-7740 x211 Direct: 612-392-6814 EFax: 952-873-7425 jlumby(a)mnvoip.com<mailto:jlumby@mnvoip.com> ________________________________ To unsubscribe from the MICE-DISCUSS list, click the following link: http://lists.iphouse.net/cgi-bin/wa?SUBED1=MICE-DISCUSS&A=1

Sept. 29, 2016

Re: EX4200 (1G Switch) Packet Loss -> 10G Participants

by Matthew Beckwell

Jeremy, Thanks for your efforts on this. Personally, i'm fine with next Tuesday afternoon. I think the general consensus was mid-day maintenance is fine, provided we send out something to MICE-ANNOUNCE (probably by the end of the day today) so people can shut down their sessions in advance if they prefer. I don't remember...did we decide if we're moving the 1G Optical folks on the existing 4200 expansion module (5 Nines, ClaimLynx, CDW/Berbee, Velocity Telephone) to the 4550? ~Matthew On Thu, Sep 29, 2016 at 8:35 AM, Jeremy Lumby <jlumby(a)mnvoip.com> wrote: > As many of you know after Tuesday's meeting, I moved my connection over to > the new 4200 that was donated by Compudyne. The topology is that the > existing three switches are connected together in a ring with Juniper > stacking cables, and then the new 4200 that I am testing has 2 10G ports > lagged to the existing switches, one of which goes to the 4500, and the > other goes to the 4550. When I initially connected, I was seeing loss to > all three of the older switches. The loss was 0.4% to the 4500, and 0.2% > to the 4200, and the 4550. I then shutdown the LAG port that ran between > the 4550, and the new 4200. The new loss numbers over a 6 hour period came > back as 0% to the 4500 and 4200, and 0.2% to the 4550. I then switched it > so that only the port to the 4550 was active in the LAG group, and the loss > changed to 0.3% to the 4500, 0.2% to the 4200, and 0% to the 4550. I > believe this clearly points to the loss being in the stacking of the > existing switches. My suggested next step would be to schedule a > maintenance window where the existing members are moved from the old 4200 > to the new one, and the LAG port from the new 4200 to the 4550 is left > shutdown. This will enable the existing 4200 users to experience less > loss, while we can see if removing the old 4200, and associated stacking > cables has removed the loss from the existing stack. Because of my limited > availability I am proposing doing the maintenance next Tuesday afternoon. > Let me know if this works, or if someone else with better availability > would rather do it. It should only briefly effect the 1G members, and not > the 10G users. I plan on moving the route servers one at a time leaving > plenty of time for BGP to re-establish prior to moving the second one. > > > Jeremy Lumby > Minnesota VoIP > 9217 17th Ave S > Suite 216 > Bloomington, MN 55425 > Main: 612-355-7740 x211 > Direct: 612-392-6814 > EFax: 952-873-7425 > jlumby(a)mnvoip.com >

Sept. 29, 2016

Re: EX4200 (1G Switch) Packet Loss -> 10G Participants

by Jeremy Lumby

As many of you know after Tuesday's meeting, I moved my connection over to the new 4200 that was donated by Compudyne. The topology is that the existing three switches are connected together in a ring with Juniper stacking cables, and then the new 4200 that I am testing has 2 10G ports lagged to the existing switches, one of which goes to the 4500, and the other goes to the 4550. When I initially connected, I was seeing loss to all three of the older switches. The loss was 0.4% to the 4500, and 0.2% to the 4200, and the 4550. I then shutdown the LAG port that ran between the 4550, and the new 4200. The new loss numbers over a 6 hour period came back as 0% to the 4500 and 4200, and 0.2% to the 4550. I then switched it so that only the port to the 4550 was active in the LAG group, and the loss changed to 0.3% to the 4500, 0.2% to the 4200, and 0% to the 4550. I believe this clearly points to the loss being in the stacking of the existing switches. My suggested next step would be to schedule a maintenance window where the existing members are moved from the old 4200 to the new one, and the LAG port from the new 4200 to the 4550 is left shutdown. This will enable the existing 4200 users to experience less loss, while we can see if removing the old 4200, and associated stacking cables has removed the loss from the existing stack. Because of my limited availability I am proposing doing the maintenance next Tuesday afternoon. Let me know if this works, or if someone else with better availability would rather do it. It should only briefly effect the 1G members, and not the 10G users. I plan on moving the route servers one at a time leaving plenty of time for BGP to re-establish prior to moving the second one. Jeremy Lumby Minnesota VoIP 9217 17th Ave S Suite 216 Bloomington, MN 55425 Main: 612-355-7740 x211 Direct: 612-392-6814 EFax: 952-873-7425 jlumby(a)mnvoip.com

Sept. 29, 2016