|
Fatal posted:Anybody else loving the wonderfully non-functional 15.0.2(SE2+)? Lets see, 15.0.2(SE2) has a memory leak (on 2960Ss) where if you have too many devices requesting DHCP you lose console access until reboot. 15.02(SE3) has a TACACS bug (on 3560/3750s) that kills all access, yaaaaayyyyy for summer deployment ie, busiest time of the year for me. My favourite 15.0.2 bug that I found is that the TenGig interfaces connected to Nexus 7009s wouldn't come up when they boot up.
|
#
?
Jun 27, 2013 17:58
|
|
|
|
| # ? May 28, 2024 11:46 |
|
|
Fatal posted:Anybody else loving the wonderfully non-functional 15.0.2(SE2+)? Lets see, 15.0.2(SE2) has a memory leak (on 2960Ss) where if you have too many devices requesting DHCP you lose console access until reboot. 15.02(SE3) has a TACACS bug (on 3560/3750s) that kills all access, yaaaaayyyyy for summer deployment ie, busiest time of the year for me. I had issues with 15.x on switches with monitor sessions going stale. If the destination port for the monitor went up/down, you had to rebuild the monitor session in order to get data flowing gain.
|
|
#
?
Jun 27, 2013 18:01
|
|
|
Does anyone have experience with purchasing dark fiber? We're in the Sunguard colo in NJ, close to NYC, and our new datacenter is going to be in a nearby Equinix NY4 datacenter - we're looking to do dark fiber between the two (less than 20 miles) and just wondering from a hardware standpoint what we would need to do to support it. We're not looking to utilize DWDM but basically just have a 10gig fiber between the two that we can manage completely.
|
|
#
?
Jun 27, 2013 19:03
|
|
|
You just need long range SMF ethernet optics. Nothing super exotic required. Will work in any 10G capable switch from any vendor. Also Equinix NY4 is a cool DC. As for Dark Fiber, if it doesn't already exist between the DCs you aren't going to be able to reasonably get it. But there are lots of managed fiber services in that area, so I'm sure you can find a dedicated layer2 10G between those places. ate shit on live tv fucked around with this message at 19:19 on Jun 27, 2013 |
|
#
?
Jun 27, 2013 19:16
|
|
|
Cool thanks, that's basically what we thought but some of the poo poo I read on mailinglists make me question myself. We're looking for the cheapest option since those 10g optics with an SLA can be pricey.
|
|
#
?
Jun 27, 2013 19:53
|
|
|
Sepist posted:Cool thanks, that's basically what we thought but some of the poo poo I read on mailinglists make me question myself. We're looking for the cheapest option since those 10g optics with an SLA can be pricey. Dark fiber won't be cheap, you'll likely want 10gbase-ZR optics for that reach.
|
|
#
?
Jun 27, 2013 21:15
|
|
|
We were quoted at $2,100 a month with a 3 year contract which is actually cheaper than the managed fiber solutions we're considering
|
|
#
?
Jun 27, 2013 21:58
|
|
|
Sepist posted:Cool thanks, that's basically what we thought but some of the poo poo I read on mailinglists make me question myself. We're looking for the cheapest option since those 10g optics with an SLA can be pricey.
|
|
#
?
Jun 27, 2013 22:23
|
|
|
Powercrazy posted:You just need long range SMF ethernet optics. Nothing super exotic required. Will work in any 10G capable switch from any vendor. I guarantee it's probably a lot cooler than NY7.  Is it the Seacaucus one? Also the blue lights get annoying after 12 hours.
|
|
#
?
Jun 28, 2013 01:41
|
|
|
Sepist posted:We were quoted at $2,100 a month with a 3 year contract which is actually cheaper than the managed fiber solutions we're considering God, that's so loving cheap. My agency still has a hard on for paying for a DS3, which should cost about the same. Actually, they probably pay more because they're the government.
|
|
#
?
Jun 28, 2013 03:24
|
|
|
psydude posted:God, that's so loving cheap. My agency still has a hard on for paying for a DS3, which should cost about the same. Actually, they probably pay more because they're the government.
|
|
#
?
Jun 28, 2013 03:36
|
|
|
I was actually surprised at how cheap dark fiber was. I want to say I was quoted 13k/mo for a redundant loop between 60 Hudson and one of the Equinix facilities in NJ. ~7k/mo for a loop between 2323 Bryan and 1950 Stemmons in Dallas. The price of intercity has seemingly crashed as well. We just upgraded our dfw-lax-jfk ring from 1Gbps to 10Gbps and saved 7k/mo. Granted the 1Gbps price was almost 5 years old but that's a pretty big drop. 1Gbps of L2 from 56 Marietta to 60 Hudson for < 2k/mo on 1yr term. Getting crazy out there.
|
|
#
?
Jun 28, 2013 03:59
|
|
|
tortilla_chip posted:Anyone attending Cisco Live in Orlando? Ixia unleashed a horrible blight on the conference with their fedora giveaways. The shortest blue haired girl at the VSS Monitoring booth was super cute, though; I definitely feigned a lot of interest in their product.
|
|
#
?
Jun 28, 2013 04:01
|
|
|
Always check the connectors on line cards before you insert them, especially if they're used.
|
|
#
?
Jun 28, 2013 04:55
|
|
|
Bluecobra posted:Why would optics with an SLA matter? Wouldn't you have redundant fiber connections between the two data centers? Good point  FatCow posted:I was actually surprised at how cheap dark fiber was. I want to say I was quoted 13k/mo for a redundant loop between 60 Hudson and one of the Equinix facilities in NJ. ~7k/mo for a loop between 2323 Bryan and 1950 Stemmons in Dallas. I think the price was hiked up due to it not being a managed service, thus the provider not making as much as managed fiber - then they realized it was too pricey and not selling well so they're not making money anyway, and it crashed. That's my take on it at least.
|
|
#
?
Jun 28, 2013 15:38
|
|
|
Absolutely. There was a large glut in Dark Fiber in the New York Metro Area after 9/11. A bunch of independent companies came to the same correct decision that many new york based companies would move their critical services out of NYC. Thus BT/Abovenet/Equinix/NTT/Verizon laid a whole bunch of fiber in that part of Jersey. When lots of companies do that you end up with a glut of fiber similar to the glut that exists in the Texas "Telecom Corridor". So there is a remarkable amount of excess fiber capacity. Another advantage of dark fiber is that if it exists there is a LOT of it. The marginal cost to lay 1000 fibers 100 miles as compared to 2 fibers is almost the same. Buying the actual fiber is the cheap part of laying fiber.
|
|
#
?
Jun 28, 2013 19:49
|
|
|
falz posted:Always check the connectors on line cards before you insert them, especially if they're used. This is probably the result of moving a chassis switch with the line cards in them. I always take all the line cards/power supplies/fan boards out before shipping/moving a chassis switch.
|
|
#
?
Jun 29, 2013 15:52
|
|
|
chestnut santabag posted:My favourite 15.0.2 bug that I found is that the TenGig interfaces connected to Nexus 7009s wouldn't come up when they boot up. Owned. I have never heard a company say as many negative things about a product as I heard Cisco say about the entire Nexus line at Cisco Live last week.
|
|
#
?
Jun 30, 2013 04:36
|
|
|
Forgive me for not putting a proper effort, but I can't find a general AGS+ hardware guide. I have forgotten what cards are what, past the CPU/MCI's and am getting a lot of 404's on cisco.com. Was hoping to get old nelly up on the internet so she can see how much has changed.  I have ethernet or FDDI, really hoping I can get the cards in the proper multibus/cbus slot and get the ethernet's connected properly. Gonna have to snag a better 8.x IOS as well.
|
|
#
?
Jul 3, 2013 00:17
|
|
|
Herv posted:Forgive me for not putting a proper effort, but I can't find a general AGS+ hardware guide. http://sobek.su/Docs/univercd/cc/td/doc/product/core/cisagspl/agscfig/43494.htm Has a diagram of the slot layout, and other install/upgrade topics.
|
|
#
?
Jul 3, 2013 00:53
|
|
|
ragzilla posted:http://sobek.su/Docs/univercd/cc/td/doc/product/core/cisagspl/agscfig/43494.htm Thanks Ragz, I appreciate it. Stay tuned.
|
|
#
?
Jul 3, 2013 01:18
|
|
|
This should be a simple question but I can't figure it out and google is being difficult. I have a standard Dual Hub and spoke topology over MPLS via GRE/IPSEC. Each remote site is peered to each hub router over a GRE Tunnel i'm using EIGRP for routing. So my question: Given a specific Neighbor or a specific interface, how can I see which routes I'm learning, but only for that specific interface/tunnel. So while "sh ip ro eigrp" will give me all eigrp routes, i want just which networks are being advertised from Tun24 or neighbor 10.1.24.2. Ideally I'd want something like: "sh ip eigrp nei x.x.x.x" Networks learned from Neighbor x.x.x.x 10.1.1.0/24 10.1.2.0/24 172.16.168.0/20
|
|
#
?
Jul 3, 2013 21:35
|
|
|
Powercrazy posted:This should be a simple question but I can't figure it out and google is being difficult. I have a standard Dual Hub and spoke topology over MPLS via GRE/IPSEC. Each remote site is peered to each hub router over a GRE Tunnel i'm using EIGRP for routing. How about "show ip ro | inc Tunnel24"? Or you could do "show ip ro | inc 10.1.24.2" since that'll be your next hop IP.
|
|
#
?
Jul 3, 2013 21:51
|
|
|
Yea that works and I'm pretty sure that is the ONLY way to do it. After discussing it a bit with a peer, it comes down to the fact that the EIGRP Process keeps no database of updates. So even though it receives an update on a specific interface, sourced from a specific neighbor, with a specific list of networks, it doesn't store that information anywhere. It acts upon that information and installs the route into the routing table. So if you want to act on any of the information the only place it exists is in the routing table. Basically there isn't such a thing as 'sh ip eigrp database.' As an aside the reason I don't like output filtering is that it is dependent on your terminal settings. So for example if you shrink your terminal to a 20x20 character square and type "sho ip ro | i Tunnel24" You will get a different output then if you had a 400x400 character terminal because of innate IOS output formatting. Also it's better to output only the information you want rather than filter on output. So for example if you want to see all the static routes on an internet edge router taking full BGP routes you should type "sh ip ro static" instead of "sh ip ro | i Static" Even though both will give you the same information. ate shit on live tv fucked around with this message at 23:53 on Jul 3, 2013 |
|
#
?
Jul 3, 2013 23:42
|
|
|
Powercrazy posted:Basically there isn't such a thing as 'sh ip eigrp database.' I don't know how much detail it provides in meshed situation (I only use EIGRP to interop with existing customer networks), but the equivalent to ospf database is "show ip eigrp topology" http://packetlife.net/blog/2010/aug/9/eigrp-feasible-successor-routes/
|
|
#
?
Jul 4, 2013 00:57
|
|
|
ragzilla posted:I don't know how much detail it provides in meshed situation (I only use EIGRP to interop with existing customer networks), but the equivalent to ospf database is "show ip eigrp topology" The problem with using the topology table is that networks and interfaces are on separate lines, e.g.: code:
|
|
#
?
Jul 4, 2013 01:17
|
|
|
Ever since I updated my Linkedin with my CCNA cert I've had three phone interviews with companies in North Carolina, Ohio, and Iowa. I don't care to move to any of these states but I was interested in what they'd offer, and it's been anywhere between 10k - 20k more than what I'm making now. So, at the end of this year, I'm going to ask that I receive more competitive pay. We've added at least an additional 100k a month in revenue. I know they can afford me. If I quit and they put a help wanted ad out for a CCNA cert, he's gonna want what I'm gonna ask for. It's better for them to keep me since I know the network, know the company, and I don't mind doing things above my pay grade. If not, I'll cash in my ESOP and walk.
|
|
#
?
Jul 5, 2013 15:04
|
|
|
Why wait until the end of the year, you're only stunting your own salary by doing so. It is really hard to find good network engineers right now, most employers know that and if you are good then you should have no problem asking for a raise.
|
|
#
?
Jul 5, 2013 16:22
|
|
|
Sepist posted:Why wait until the end of the year, you're only stunting your own salary by doing so. It is really hard to find good network engineers right now, most employers know that and if you are good then you should have no problem asking for a raise. Can't cash out my ESOP till the end of the year. I don't want to give them an ultimatum without the financial safety net in case they actually believe that paying a kid who is straight out of Community College 30k a year is a smart decision to make. "Hey Billy, we need you to setup an internal BGP peer with our new Google caching service. Get a MOP to me by tomorrow!"
|
|
#
?
Jul 5, 2013 16:33
|
|
|
Psh, move to NC and work for me. I need another network engineer badly.
|
|
#
?
Jul 5, 2013 20:21
|
|
|
FatCow posted:Psh, move to NC and work for me. I need another network engineer badly. ...oh do you now?
|
|
#
?
Jul 5, 2013 20:25
|
|
|
I need to span numerous vlans across 2 datacenters. Will EoMPLS work for me? 6500's with Sup720's on 12.2SXJ5. Thanks in advance.
|
|
#
?
Jul 5, 2013 22:35
|
|
|
World z0r Z posted:I need to span numerous vlans across 2 datacenters. Will EoMPLS work for me? 6500's with Sup720's on 12.2SXJ5.
|
|
#
?
Jul 5, 2013 23:08
|
|
|
World z0r Z posted:I need to span numerous vlans across 2 datacenters. Will EoMPLS work for me? 6500's with Sup720's on 12.2SXJ5. Let me guess, vmware? I had to fight so very, very hard to get my system people to understand why I wasn't interested in extending a vlan between bicoastal datacenters.
|
|
#
?
Jul 5, 2013 23:40
|
|
|
adorai posted:I would suggest that you also look at straight Ethernet products. They are pretty cheap. We have a few 100mbps Ethernet links of 100+ miles and pay about $1000 per endpoint. They are configured in a standard any to any Ethernet network. Another local provider that was built with federal grant money has quoted us $680 per endpoint per YEAR for gigabit TLS, but we have to perform our own fiber build. Which isn't all that bad, since they cover most of our footprint pretty well anyway. I am pretty sure he isn't talking about setting up a /30 and using an Ethernet point-to-point connection to route between sites (which is the right thing to do).
|
|
#
?
Jul 5, 2013 23:49
|
|
|
World z0r Z posted:I need to span numerous vlans across 2 datacenters. Will EoMPLS work for me? 6500's with Sup720's on 12.2SXJ5. Yes EoMPLS will work, but try moving the inter-dc network to layer 2 first. I use EoMPLS in my own network and while it works it's inefficient and a pain to manage so i'm turfing it.
|
|
#
?
Jul 6, 2013 00:29
|
|
|
I have no budget for new tech so OTV on N7k is out; I can't do VPLS because no ES+ or SIP mod. I need to span vlans to facilitate a migration without any IP readdressing. We'll leak host routes at the new sites bit by bit for IP routing. I know it won't scale and I don't care. I can't spend any money and I have to move things from one DC to the other with minimal impact.
|
|
#
?
Jul 6, 2013 01:07
|
|
|
jwh posted:Let me guess, vmware? Luckily for me the two DC's are in the same metro area. 3ms between sites.
|
|
#
?
Jul 6, 2013 01:10
|
|
|
jwh posted:Let me guess, vmware? B-but you don't understand, we have a critical need to vmotion VMs between the data centers! Vmotion has made server/application guys so lazy when it comes to building redundancy  .World z0r Z posted:Luckily for me the two DC's are in the same metro area. 3ms between sites. Surely you can get a dark fibre service for less than a managed service if they're only 3ms apart?
|
|
#
?
Jul 6, 2013 04:02
|
|
|
|
| # ? May 28, 2024 11:46 |
|
|
ruro posted:Surely you can get a dark fibre service for less than a managed service if they're only 3ms apart? Not under the constraints I have.
|
|
#
?
Jul 6, 2013 04:14
|
|