Otanx,

on the cat 6509e with quad-sup - the failed supervisor dies and the standby comes online after the entire chasis reloads - it provides a deterministic outage window measured in minutes and you are back online. Replacing failed supervisor can be done without any priming or commands.

on the cat 6509e with quad-sup - the plan is later this year to have SSO within the chassis so can do upgrades hitless.

on the nexus 7009 with quad-sup - the failed supervisor dies and standby comes online without any reloading -completely hitless. it also already allows for hitless upgrades.

The bandwidth performance on the Nexus 7009 really does seem to be more than we need or will ever use - but at same time - I do not want to recommend and purchase a 6509e that will be announced end of sale or discontinued within 4-5 years - our partner says the 4500e/6500e should continue to be sold until 2020. I understand Cisco Live is in a couple months and maybe the 7009 will even be upgraded with newer sup's and line cards, and make the 7009 purchase look bad timed...so much to decide on...

- our partner says the 4500e/6500e should continue to be sold until 2020.

partner is wrong 4500E were just announced EOL yesterday, end of sale date April 16, 2013

see thread in cisco general forum
viewtopic.php?f=36&t=30613

partner is wrong 4500E were just announced EOL yesterday, end of sale date April 16, 2013

see thread in cisco general forum
viewtopic.php?f=36&t=30613

... I do not want to recommend and purchase a 6509e that will be announced end of sale or discontinued within 4-5 years - our partner says the 4500e/6500e should continue to be sold until 2020.

no doubt they'll do the same as they did for the 4500E
EoL the 6500-E and replace it with a 6500+E
money grubbing grubbers....

FYI, Cisco gave us a product update today and the 6500 is going to be around for quite awhile. They have some solid features coming in the next year.

FYI, Cisco gave us a product update today and the 6500 is going to be around for quite awhile. They have some solid features coming in the next year.

So does that change your opinion about not deploying 6500s or are you seeing new life in them?

So does that change your opinion about not deploying 6500s or are you seeing new life in them?

If the 6500s could compete at 10g I'd probably back off the Nexus for a year or so.

I'll put down a few words, so feel free to discard them at will.

We've implemented a similar solution with two 7009 switches with F2 modules, 5596UP switches and 2232PP FEXes, and there are a few things to be aware of with regards to the 7k platform in general and the F2 modules in particular.

1) The F2 mainly a High Troughput line card, whereas the M series is for Features. This means that you need to carefully consider what you want to do and figure out if the F2 will do it for you. For example the data sheet says it does NetFlow whereas the latest software specifically says it _doesn't_ do it on the F2. TAC says to wait for 6.1 release in Q2Y12.

2) IPv6 on NX-OS (caveat: on 7009 chassis with F2 module) is missing a few features such as RA priority and suchlike. Not sure if this is coming or not, but it's a bit of a bummer if you wanted to make one chassis the primary unit for a particular VLAN (both v4 and v6).

3) We're currently engaged in a TAC case concerning some problems we're having with HSRP and DHCP relaying - the DHCP discovery is forwarded to the server, but in some cases the reply doesn't reach the client. In this particular case the servers subnets are routed on both 7009 switches but the primary switch for the client VLAN might not be the primary for the servers, causing a degree of asynchronous routing (long story). At the moment it looks like a ARP or layer 2 issue but we're not sure yet.

4) NX-OS is very much into the new NetConf management way which means many of your old SNMP tools won't work (this goes for both 7k and 5k). DCNM is not a very mature product as far as I'm concerned and leaves a lot to be desired. If you have SNMP tools you need to make sure you can use them with the new platform - the MIB support on the 5k is laughable, for example. If you do everything by CLI and spurn fancy GUIs as toys for children you're golden

There have been a few other minor niggles, such as the mgmt0 interface on two 5k switches actually changing MAC addresses after a software upgrade (which meant the NMS thought they were down for a good 30 minutes until we could investigate - the MAC addresses had been incremented by 1 for some reason). They also had a _serious_ bug in the 5.0(something) version of code for the 5k that made the vPC pretty much useless. This was fixed in the 5.0(something + ohcrapIcantbelievewescrewedthisup) release.

I'm not saying to steer clear of the 7k/5k switches, just be aware of what your expectations are and make sure your kit will deliver (ideally with close support from your VAR, preferably with documented deliverables!). The Nexus series appears to be the way forward, but it's not the trusty 6500 platform of yore quite yet. Granted, the 6500 has been allowed to mature for a long time (some would say too long), but there are a few issues on the Nexuses that are not worthy of a Cisco flag ship product no matter how you slice it.

In short, the 7k/F2 combo is a good product with a lot going for it. In a year or two with loving software updates it may even become great. As it stands I feel a tad cheated by Cisco SlideWare and the SE/VAR, but as we've got a sunset/horizon of about 10 years for this kit it might have improved significantly by the time it gets scrapped in favour of something Greater(tm) and we start all over again

I apologise if I hijacked the thread, but it's been a full day of smashing my face against the wall of "don't tell me this shiny new Nexus actually fails this badly at something that shouldn't even be an issue"

If the 6500 did 550gig/slot, I would personally throw my 7Ks away, as would pretty much everyone. But we all know it will never compete at 10gig.

How many 10GE ports can be used with proper subscription based on what you have in the Aggregation / Distribution / Access layer with two 6509? 6 line cards remaining per chassis if 2 supervisors and a services module is installed.

I'm thinking about 4 Nexus 5596UP's - at the server aggregation layer doing local L2 switching - with 4 10GB uplinks to core per 5596 and 4-6 Catalyst 4507's having one 10GB uplink to each core per 4507.

Seems if go with the 7009 we will have 5 line cards empty, or 4 cards if we add another 48port 10gb which would equal adding two fully loaded 6509e line card capacity?

What am I missing? Maybe I am looking at this with not a long enough timeframe?