As stated previously in order to try and pass the SP Lab the 2nd time around I went into real post-mortem analysis of my lab attempt and the training materials available. I also have gone through multiple Cisco Presentations, NANOG sessions, Networkers aka Live, listened to various podcasts, videos on demand and finally delved deep into Cisco Forums. Key points uncovered and reasons for failing the SP Lab.
- Over-confidence in pre-configurations – you must verify everything before you start – I cannot over emphasise this point.
- Ensure your lab matches your topology – ping directly connected ethernet devices, L2, some basic L3, some MPLS and some VRF are already pre-configured – check this pre-config. Ref: Networkers 2009 CCIE SP Techtorial
- INE SP Mantra – Build and Verify then move on!!!!
- Running out of time due to over-configuration, IOS Bugs, router reloads, misreading of the exam questions, troubleshooting issues, etc
- Staying with IOS Bugs, certain SP stuff works one way on 2600’s and another on 3600’s – be aware of the IOS nuances, reload at least twice during the lab.
- Sequence is everything in the SP exam – e.g. MPLS TE – you must enable MPLS TE capability for the IGP’s prior to configuring the MPLS TE itself or it might not work Ref: INE CCIE SP CoD. It does not mean it won’t work but it might not work. Also it may take up to 20 minutes to decide to work. This is an issue close to my heart as I configured MPLS TE perfectly on my lab or so I thought but no marks.
- Be aware that it appears the proctor has the ability to remote onto your PC – this does not mean they do anything other than observe.
- Ensure your IGP & EGP are completed and stable prior to starting multicast section to ensure you do not encounter SPF issues. Ref: INE RS V4
- Know the nuances regarding reachability in the SP Lab as against the RS Lab – TCL alone here will not suffice, remember ping , ping vrf, etc
- Also remember that full reachability is not necessarily required as per SP networks in general Ref: CCIE SP ‘Ask the Expert’
- Verification is unbelievably important [Stressing the obvious] – know your commands, know what to expect form the outputs – e.g. MPLS see an outgoing lable be an incoming label on the next hop
- Use notepad for large configurations or even for troubleshooting – it’s harder to scroll page after page on SecureCRT on a per router basis than to ‘select all’ copy to notepad and compare two router configurations that way
- Answer what is explicitly asked for and nothing else – no points lost for over-configuration
- Do not reload the routers before you leave the exam Ref: CCIE SP Tips – Cisco Learning Network
- Watch out for label forwarding breaks in the MPLS VPN’s where VPN routes will be able to exchange between PE’s but VPN traffic cannot be delivered Ref: CCIE SP Tips – Cisco Learning Network
The title of this blog entry is ‘Is the fog lifting’ – I am beginning to see the various nuances now – e.g. make sure you match internal external on your OSPF redistribution as external routes not passed by default, local preference BGP attribute is reflected throughout AS as against other attributes such as weight which is locally significant, IP addressed removed when enabling VRF forwarding on interface because the prefix is taken from the IPv4 table and inserted into the VRF table, etc.