Things to do at work on your first day

[ohforce55]

Hi,

I always wonder how to analyze the company network as you just start at the new environment.

I have done tracing all the cables from core switches so what is connect to what as wells as figured out the serial numbers of each devices.

Besides that, in order to troubleshoot better and for your benefits,

what else do I have to look? How and what should I analyze? How do you know how the traffic flows? etc..

What commands do I have to use to analyze the network in a perfect way?

I had my boss (senior engineer) who analyzed the company network for 2 weeks since he started.

Since I'm not that level, I'm really curious what needs to be analyzed and how..

I will also very appreciate if you could give me any other tips to look at.

Thanks in advance!

[Georg Pauwen]

Hello,

actually, one of the first things I would inquire about is how users experience the network. The network is for them, you work for them, so ask them if performance, speed, reliability, and the overall impression are acceptable, and what can be improved...

[#3]

Oh dear, you're opening a can of worms with this one... "so ask them if performance, speed," lol

Every end user you talk to will say it's too slow.

Okay, enough of my pessimistic approach. haha I would agree with finding all documentation related to the infrastructure. Instead of tracing cables, you can view what's connected to what with "sh cdp neigh" if you're looking for a physical topology. If you're looking for a logical topology, I suggest looking at the OSPF database (assuming you're using OSPF) and building the logical topology of your network on a sheet of paper. It's an AWESOME way to really learn in depth OSPF. If there is redundancy, try to figure out WHY traffic is using a preferred path. Look at spanning tree and see where your root bridge is and WHY it's your root bridge. One thing to definitely NOT do is stress about it. Take it slow and learn it at a deeper level. Depending on how complex the network is, it could take you a month or more to fully understand traffic flow. If you have another net admin/engineer, see if they will mentor you in any way and ask questions when you cannot figure something out.

Good luck and have fun!!!!

[ohforce55]

Thank you for your reply. Besides looking into routing protocols, what other things do I have to look? 

Except 'sh cdp ne d, sh ip route, sh ip eigrp etc...', are there any other commands that you usually use to analyze the network of your environment?

I still don't understand how I can figure out how traffic flows.

Only one that I can think of is just use traceroute command from my pc or core switch but what else method can I use in order to figure it out in a certain way?

[#5]

I would suggest practicing in MS Visio drawing diagrams. If you're starting from scratch and nothing is documented, your manager will love you if you build a network diagram.

Look at layer 2 paths. Find the root bridges and understand STP, RSTP, backbone fast, and uplinkfast.

What is the IGP your company uses? I feel I'm pretty darn good with OSPF. Traceroute is very good to look at the path used by your LAN. After you see this path, you can look at the devices and find out WHY this path is chosen.

I'm assuming Cisco vendor for this: Look at sh cdp neigh. This can show you what is connected to core devices and you can build out from there. If you want to build an OSPF logical topology, you can use "sh ip ospf data". This will show you the LSA database and you can build the topology from there. I find that kind of fun. lol I'm a dork. I just really enjoy networking and love to continue to learn and do better. I also have an amazing mentor at my job.

Good luck and I'll keep an eye on this post! If there is something I can answer, I'll be happy to!!!

Mike

[ohforce55]

Thank you for your reply.

Yeah, I have using MS visio diagram. We are using EIGRP.

I have another questions. Besides analyzing, what else do I have to check?

How can I check if the status of running devices (I use these commands 'sh process cpu his' and 'sh log'. What else do I need to see?)

How can I check if the connections are okay? Is there any way to check if IGP or BGP connections or the routes anything that is in route table or arp table etc are running okay?

I weekly check the logs and cpu utilization and do the back up configurations. What else do I have to do in a period of time?

Please let me know. Thank you!

What to check? device? connection well?
sh log

daily basis

besides analyzing network
anything else that has to be looked?

[dperezoquendo]

Hello,

I've  been more on the Engineering side of the house as opposed to the admin but as all the others here had basically suggested, if there is no documentation, start creating it.

By building detailed network diagrams, you will not only begin to better understand network design and the overall structure of the environment you're supporting, but you will also most likely identify key concerns that you will likely investigate. If these concerns are valid and can improve the security and performance of the network, you want to bring them up.

Some basic items I can think of on the top of my head would be the following:

• Switching & Routing  -
• Does your STP configs meet best practices? e.g. rapid-pvst, bpduguard on portfast enabled ports, etc
• EIGRP - Are passive-interfaces used to control adjacencies? Are you using authentication?
• Redundancy - What is the preferred path and why? Maybe see if you can schedule outages to verify redundancy works as intended?
• Security -
• Do your devices meet best security practices? E.g. port-security, not using vlan1, authentication, etc
• ACLs and Firewall rules - What are the ACLs used for? Can you improve your firewall ruleset? (e.g. cleanup redundant and unused policies)

Eh, that's all I can think of right now. In other words, I would recommend to do some research into best practices and design guides. If you see any concerns within the network, you should ask or try to figure out why that portion of the network is configured the way it is.

Also it seems your environment doesn't have any health monitoring tools emplaced (since your doing show process cpu hist), you can try checking out tools like Solarwinds or Intermapper.

Again, as others have mentioned, take your time with this. It is a lot of information to process and learn. It appears to me you were given a good opportunity to excel and grow within your current environment. Don't do the bare minimum that so many others seem comfortable with. :)

[ohforce55]

So much appreciated!!!

Can I still reply whenever something to ask pops up in my head?

Thank you so much!

[ohforce55]

Thanks for your reply.

[Julio E. Moisa]

It was a pleasure,

have a great day

:-)

[Julio E. Moisa]

Hi

I think one of the first things to do is collect all the possible documentation related to the network (infrastruture), diagrams, IP plans, routing protocols, POP's, etc. Try to get familiar with the infrastructure. 

Once the information is gathered you can start to analyze the information and configurations. And see what you can improve. Also as Georg mentioned you could talk with the users to know their experiences with the network, if they have had outages or something like that.

:-)

[ohforce55]

Thank you for your reply. What if I have to start from the scratch? If there is no any network document, where should I start from?

[Julio E. Moisa]

Hi

I think you can begin identifying the core devices and then verify the devices connected to them, logical and physically. Try to identify if the company is using a model design, like collapsed layered model or three layered model. 

:-)

[ohforce55]

Thank you!

[Julio E. Moisa]

It was a pleasure,

have a great day

:-)