The collected works of ipaddict

The big question is whether they will be able to nudge Sun aside enough to make any money on Optys. Sun already has a solid foot in the door, and is selling at very competitive price points.

SGI+AMD would be nice, but I'm not so sure it will save them from the hole they have dug themselves. Does SGI still have enough mental strength to engineer a competitive solution? After their recent bout of layoffs, my thinking is not. :(
jfeldt wrote: I have a Linksys WAP54G that I have been using as an access point client. It has been totally stable for as long as I have owned it (~8 months). Whatever you end up getting, make sure it can behave as a client also, since I think that isn't a given, as one would believe.


That is referred to as a "bridge." :wink:
Thank you jan-jaap! I am currently dumping everything onto my O200 and converting the PDFs to plaintext for easy greping. :D

Any idea how large this is going to get, and how often SGI really updates their doc tree? I wonder if I did an indepedant dump next week, how large the diff would be? Hmmm...

[Edit]Turned out to be just shy of 2.5GB. *LOTS* of 404s, though.[/Edit]
Heh. A "real" job.

I would suggest the Sybex book by Todd Lammle. It really is the best text on the market for beginners. You shouldn't really be memorizing IOS commands and subnetting, you should be learning how Cisco tends to build IOS commands and you should be able to subnet in your sleep. On the Cisco exams, believe it or not, practical experience pays better than memorization and reading lengthy descriptions of various protocols and their behavior.

You'll need to be able to configure single-area OSPF, (E)IGRP and RIP (no BGP, IS-IS or multi-area routing).

Have a decent understanding of WAN technologies and which are packet, circuit and cell switched, as well as how to troubleshoot a serial (DS*) and frame relay interface.

You'll need to be able to identify which switch in a stack is the root switch, how it got elected and which ports are blocking, as well as how STP determines which ports *to* block.

You'll need to be able to configure and troubleshoot basic VLANs and dot1q trunks (no ISL).

Be able to quickly troubleshoot basic routing issues based upon sh ip int, sh run and sh ip routing commands.

Passing is 875/1000, and the possible range is 300-1000, you'll have 90 minutes to take the test, with 2-3 sims/simlets that will take up most of your time.

If you have any (more specific) questions, feel free to ask me via PM or email. Hope the brain dump helps. :wink:
Hint from the Pros:

Named access lists *can* be edited line-by-line in the shell. Each line is assigned a number starting at 10 and incrementing by 10 for each line. When adding/removing specific access-list entries, specify the sequence number:

Access list:
ip access-list extended INBOUND
permit tcp any host eq www
permit tcp any any www
deny ip any any log

sh ip access-lists
Extended IP access list INBOUND
10 permit tcp any host eq www (0 matches)
20 permit tcp any any www (0 matches)
30 deny ip any any log (0 matches)

To remove an entry, simpy:
conf t
ip access-list ext INBOUND
no 10

Adding a new entry is just as trivial:
conf t
ip access-list ext INBOUND
15 permit tcp any host eq www
Yeah, I know. Most Cisco texts will neglect that particular feature. Why? I don't know. :?

I didn't take my CCNA until after about 6 years of working with Cisco devices day-to-day. I'm now working on my CCNP (two more tests! woohoo!), and aspiring toward CCIE status.

After I really got involved in Cisco work, I decided to purchase some equipment for a lab environment (much like I do with nearly everything else I do... I prefer hands-on experience to a paper trail), and have very much enjoyed doing work at the lower layers of the protocol stack. It's very refreshing to get through a day without hearing about some new Windows bug, or some such garbage (not that it effected me when I was admining UNIX boxen, but still...).

One note that many books will also neglect: the CCNA no longer involves *any* 1900 series switch work or 2500 series routers. Most books will cover the 1900/2500 as they were the de facto switch/router combo for the CCNA's simlets. That is no longer the case.

A 2924 or 2950 switch and a 2600 series router is recommended for reference, now.
Thanks Wolves! I feel the same way; practical, applicable knowledge is nearly always more valuable to me than the ability to pass a test, but I'm always up for a challenge! I've found the Cisco exams to be a pretty good indicator of one's field knowledge, though. They seem to hit the mark closer than other trade certs.
Something like that. Typically in a simlet, you'll get a partially setup network or a network with some issue to troubleshoot. You'll be given consoles into the appropriate network devices, and some addressing information. You can check out Cisco's site for some practice exams (they're flash-based, you're forewarned).
hamei wrote:
looks like there's some people here with Cisco hands-on, so mine the knowledgebase

The nick isn't coincidence... :wink:

hamei wrote:
It's time to upgrade the network here. Pix would be nice but they aren't gonna go for that. Maybe next year.

Well, when you are: a 506E is a nice little small-office, VoIP-capable PIX. The price really isn't bad, either. If you need something a bit heavier duty, the quintessential 515E works great and the ASAs aren't really presenting any compelling reasons to upgrade off of PIX hardware, yet.

hamei wrote:
There must be a worse method out there but I haven't heard of it.

You're right. That's disgusting. But alas, it isn't the only similar case I've heard of. :roll:

hamei wrote:
Anyway, a 2900-series switch should be fine

Yup. A 2924/48 or 2960 (preferred) provides a good access-layer foundation for growth. Eventually, you may add another 2900 series, and maybe a 3550/3560/3570 to do VLAN routing and QoS tagging for you. Especially, with VoIP, a layered approach will be more robust and flexible in the end. However, the 2900 is a fine switch to begin with.

hamei wrote:
10/100 fast ethernet. One day tho, I'd like to kick Skype into the trash and get real VOIP. Also, it would be nice to eventually replace the el-crappo optic-to-ethernet transceiver with a medium-length optical WAN connection right into the router ? For wireless we can live with the Dropped-Link subnet ... keeping economics and future expandability in mind, what would you experienced Cisco admins recommend ?

Well, everyone has FastEthernet on their routers. Really, that isn't a hard requirement to meet. Even the "lower tier" Cisco kit is getting GigE now (2800 Series). The 2800 would be perfect for you. The 2811, for example, can have 2 built-in GigE ports, it can have L2/3/4 switch modules added as the need may arise, WLAN Management interfaces, a GigE SFP (fiber) HWIC for WAN access, and all the Call Manager integration you could ask for. Note the 2801 won't suit your application, but any of the other 2800 series would do just fine and grow with your company.
chervarium wrote:
You should not come to me on an interview with a CCNA only as I willnot hire you.

What a cheery greeting. Are all Bulgarians so welcoming? Perhaps the next time you feel like saying something in this veign you could provide some useful feedback for those individuals who may be interested? That would be undoubtedly be appreciated more than a verbal door-slam. :roll:
hamei wrote:
It would, but I think you missed the part about "cheap". Lowest price I can find for a usable 2800 is about $1500. :-( That's why I was kinda looking thru the 2600 -3600 serieses. A 1721 looks like it would be adequate but seems to have some significant shortcomings.

In Cisco-land $1500 *is* cheap. :wink: Of course, that's just the bare chassis, you would still need the HWIC for fiber.

With anything below a 2691/3660, you're still going to need that transceiver, I think. The NM-1GE (Gigabit SFP NM) is only supported on the higher-end models in those series that have the backplane capacity to support fiber.

The voice capabilities are roughly equal (ignoring trunk capacity and backplane speed) on 17/26/3600 series, as they will all use the same WICs and the 26/3600s will both use most of the same NMs.

You won't get much wireless integration capability out of a 17/26/3600, at least not on the level of the 28/37/3800 series, but you could do basic inter-VLAN routing, and given a RADIUS server, dynamic policy assignment.

Forget the 1721. It's a broadband router, like your D-Link; the cost-benefit ratio is way too high.
I have the Boson software, and it doesn't require Internet connectivity. It requires Internet *registration*, but no actual connectivity is required.

I put it on my SO's AutoCAD PC to try it out and didn't find anything in it that I couldn't do with my "real" routers. I'd be willing to relinquish the license to you, if you'd like - it's of no use to me. PM me if you're interested.
ASP.NET or "classic" ASP? Here's the differences, as they'll affect your learning vector:

"Classic" ASP uses VBScript (A VB-like scripting language - go figure!) as it's primary development language. It does not make use of any of the .NET runtime, libraries or fancy-pance JIT "stuff."

ASP.NET can use any of the .NET languages and often makes use of things like XML, SOAP and WSDL.

While I'm certainly no expert on either of these topics, I have done a bit of maintenance work with both. Nothing I really enjoyed or was amazed by though.

If you're going .NET, I would suggest starting with C#. You'll probably find it most familiar.
WolvesOfTheNight wrote:
Well, I took the test today and passed it!

Was there ever any doubt? :wink:

Let me know when you start up on the CCNP - I'm sitting for the fourth exam in less than a month and after that I might have some equipment to "liquidate." As always, drop a line if you have any questions.

WolvesOfTheNight wrote:
How many people do you know that are installing ISDN on new routers so that they can use it as a back up network connection?

Actually, this is just a hint of things to come. CCNP and CCIE Voice cover ISDN in much greater depth. The coverage is CCNA is entirely to provide a foundation for Voice/VoIP coverage in later certs.

With regard to Cisco's "stagnation", I think that is more a lack of time spent around larger networks. Most ISPs and the vast majority of large corporate networks run on at least a Cisco foundation, if not all Cisco equipment.

HP really is the only competitor that I've noticed making headway, and that's been only at Universities that were offered amazing prices.

I mostly see Juniper/Extreme/Foundry in (very) large ISPs, and only at the core where a Cisco 6500/12k would otherwise be.

At the edge, in the wireless, firewall and VoIP arena, there is an amazing amount of diversity. I see a bit of 3Com, Nokia, Fortinet, Juniper, Shoretel, Polycom, Nortel etc. sprinkled about.

For the most part though, it's Cisco. PIX/ASA, 3600-7500 routers, 6500/12k core switches, 3500/3700 distribution and 2900 at the access layer.
WolvesOfTheNight wrote:
However, as ISDN is now (speed & price) I think that they should just give up on it.

Again, they aren't pushing it for *data*, but many people still use ISDN for voice, where a T1 is impractical or too expensive.

WolvesOfTheNight wrote:
-One of the pervious big reasons to buy cisco was that they were multiprotocol. Now most everyone is happy with IP; who cares if your router will handle appletalk or not?

The multiprotocol argument is a fair one. All vendors implement compatible IP stacks. On this issue it's mostly a matter of price (e.g. the SOHO router).

WolvesOfTheNight wrote:
-For ethernet routers/firewalls/multilayer switches you can put a pile of ethernet cards in a linux box. I don't know how this really compares with using cisco products, but I do know of people that do it because it is cheaper.

First, in my and most other professional's opinions I'll wager, host-based routers are a joke in the enterprise space and nothing more than toys in SOHOs. They are the domain of hobbiests that enjoy doing new and interesting things with their OS of choice (which is not at all a bad thing). However, not only is there no support on such "devices" (a *huge* deal for a business), they are a kludgy mish-mash of applications and configurations.

IOS is standard across all routers that use it. The syntax is the same, upgrade paths are the same or very similar, and it's performance on various devices is nearly a known quantity. The variety of ways to configure/upgrade a host-based router of *any* kind just doesn't lend itself well to environments where the admin that built it likely will not be around when there is a problem and device performance/reliability is paramount.

WolvesOfTheNight wrote:
-Judging from what I covered in the CCNA certification, they don't seem interested in making their stuff easier to administrate. While this makes it easier for me to get a job, it is a poor long term business practice.

There is an http(s) server on nearly every Cisco device with a GUI, if you prefer that to the command line. Not all configuration options will ever be available through the GUI, though (that would lead to something on the order of the Windows Registry in size, complexity and probably instability). You can perform most common tasks and get a snapshot of the device via this method. Despite the availability of the GUI, I have never seen an admin I respected use it, and probably never will - the console is standard and predictable across devices (once you're used to it).

WolvesOfTheNight wrote:
In the past 5 years how much new & exciting stuff has cisco come out with, and how much have they dropped the prices on the old stuff. Are they coming up with new stuff that you really want in your network, or do they depend on selling retreads of old products?

Most of the development in the networking world is in wireless and VoIP right now. In that regard, Cisco has created entirely new product lines and invested in making them reliable enough for enterprise use (they run on their own equipment - believe it or not, including CallManager... ugh). Their products are never perfect (whose are?), but they are rarely behind the curve with reliable(!) hardware and the support to back it up. In the industry, I don't think anybody has anything like TAC (they'll write the config for you, if you have a service contract - most Cisco customers do whether they need it or not).

Clearly, I am behind Cisco in much the same way as some sysadmins are with Microsoft. While some would disagree with this standpoint, Cisco products have always been good to me, and when they haven't (hardware failure), Cisco has always replaced the part as fast as I asked for it. Their devices perform well, are reliable in ways many other companies can only dream of, and are a de-facto standard both within their own company and across the industry. There really is very little reason *not* to buy Cisco, if you can afford it (this is a large point of contention, even among Cisco supporters), and *no* reason for a netadmin worth their salt to not be familiar with IOS.
Honestly, I've always found the Cisco texts to be far too theoretical, especially for reference purposes. I do, however, keep a copy of the latest edition of the O'Reilly Cisco IOS in a Nutshell on my bookshelf. I found it handy when I didn't have access to the 'net (say, at a customer's site) and I needed to figure out the proper usage of a command. I've become decreasingly dependant upon it and for command reference as I've gained experience (which I think is a "good" thing).

The only other really good text of CCNA level that I'm familiar enough with to recommend is the Sybex CCNA Study Guide by Todd Lammle.

Lammle has a great writing style and a fantastic ability to cover enough of the topic areas to be applicable to the CCNA, while still allowing for the student's growth by referring to specific items (esp. with EIGRP and OSPF) that will be expanded upon in future studies/practice.

He, et al., also published an all-in-one CCNP book that I found really quite good.

I agree entirely with you with respect to introductory courses and "glossing" over topics, and I think you're spot on with regard to Cisco's CCNA texts. However, they advertise the CCNA as an "apprentice" level certification, wherein the student should learn a broad range of topics, but very little depth within any given area. This is mostly to introduce the student to technologies and ideas that they probably have had very little exposure to (e.g. link-state routing protocols, frame relay and ATM).

The CCNP goes much further in-depth on several of these topics, providing enough knowledge for the student to feel comfortable with advanced configuration and be able to choose which CCIE topic is of most interest to them.
dj wrote:
foetz wrote:

Found that, thanks :) .

Guess I just wanted to point out that the description says "no dependencies".

That seems to be a common issue. I've come across a few tardists on that list that were like that, too. It's easiest if you just nekosync up and install from a common inst directory (if you have the disk space).
Thanks in no small part to Alvers tardist building guide, I've built my first tardist! :D

It's a build of RubyGems 0.9.0. Is there somewhere I can upload it for you guys to test?

Between Alver's guide and Joerg's advice, I've got the first of hopefully many tardists built and up in /beta. Give it a shot and let me know what breaks. :wink:

In the past, I've simply built RubyGems and managed it in /opt, but this (rather trivial) build seems to be working just fine out of /usr/nekoware on my Origin and my Octane2.

Thanks Joerg and Alver!
hamei wrote:
ipaddict wrote:
Give it a shot and let me know what breaks. :wink:

I get a conflict :

Also install neko_ruby.sw.eoe (2 -2147483647) and neko_ruby.sw.lib (2 - 2147483647) from an additional distribution -- insert another CD or specify another software distribution.

Hmm... do you have neko_ruby installed?
hamei wrote:
ipaddict wrote:
Hmm... do you have neko_ruby installed?

Thought so but maybe it's b0rken :-( I had some "issues" with a famous product a while back, some things got smunched. Will have to check next weekend, time to head back to prison.

It may not be you that's borked - it is my first package. :-/
hamei wrote:
joerg wrote: Hamei have to install or update his neko_ruby.

You're right, mine was busted. But since I never use it I may just remove the thing ...

Such a beautiful language, too. :cry:

I won't hold it against you. :wink:
hamei wrote:
ipaddict wrote: Such a beautiful language, too. :cry:

Freudian sloop - just had the Assistant do a google images for "ace bandage". Now, you'd think that would be a nice clean topic, wouldn't you ?

Before the Internet got to it, I'm sure it was a fine topic. :wink:
Wow. Nice shot! :shock:
Spidy wrote: Is this similar to the ftp thing, where people scan for open ports?

Yes, it is the same thing.
Quick FYI:

I just (re-)installed neko_gimp from /current and I noticed that it (still) lacks the dependency on libart in the tardist. I thought I recalled someone posting about this issue in the past, but I couldn't find it.
nekonoko wrote:
ipaddict wrote: Quick FYI:

I just (re-)installed neko_gimp from /current and I noticed that it (still) lacks the dependency on libart in the tardist. I thought I recalled someone posting about this issue in the past, but I couldn't find it.

I don't think anyone's touched that package in a long time. Volunteers are welcome :)

That's probably why it's been a problem nearly every time I've installed it; of course, I never remember the dependency issue until I install on a clean disk and get the ld errors. I'll get a new tardist built and in /incoming this afternoon sometime.
nekonoko wrote:
ipaddict wrote: That's probably why it's been a problem nearly every time I've installed it; of course, I never remember the dependency issue until I install on a clean disk and get the ld errors. I'll get a new tardist built and in /incoming this afternoon sometime.

I had some time to tackle it this evening; did a minor version bump to 2.2.13 (current stable) and updated the dependancy list. It's now in /beta.

Thank neko. My apologies for slacking on this one.
Therion wrote:
...but it crashes immediadly to me if I try to connect with my jabberd thru TLS.

This is a known issue, IIRC.
regan_russell wrote: It is a little known fact the Alphas running Linux can run OSF/1 binaries by copying the link loader and a whole bunch of share objects across.

FreeBSD ships with an OSF/1 ABI layer that is designed for this sort of thing, as well. BSD is also generally well supported on the Alpha.
jan-jaap wrote:
ipaddict wrote: FreeBSD ships with an OSF/1 ABI layer that is designed for this sort of thing, as well. BSD is also generally well supported on the Alpha.

FreeBSD is dropping the Alpha at the next major release.

That doesn't change it's current status, and certainly doesn't prevent one from running a well-supported release well into the future. There are plenty of 4.11 boxen still serving their users well.
Nice improvement on the system drive. I'm not sure you could get much more out of that.
:shock: Nice numbers! Let's see 600 now, eh? :wink:
*Very* cool. Thanks!