- NANOG - lists.nanog.org

AT&T Noc?
by alex＠nac.net 03 Nov '98

03 Nov '98

I hate to use the NANOG list as a phone directory, but surprisingly, AT&T's contact information in the whois returns lots of voicemail and disconnections. I need a NOC contact there; please either email me at alex(a)nac.net or call me at 973.442.3401. -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- Atheism is a non-prophet organization. I route, therefore I am. Alex Rubenstein, alex(a)nac.net, KC2BUO, ISP/C Charter Member Father of the Network and Head Bottle-Washer Net Access Corporation, 9 Mt. Pleasant Tpk., Denville, NJ 07834 Don't choose a spineless ISP; we have more backbone! http://www.nac.net -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --

1 0

machine specs for Netflow collector & Analyzer
by Roy Hockett 03 Nov '98

03 Nov '98

I know this is not in the spirit of this list, however, I am sure those who have addressed this issue are participants in this list. If you have web resources, or FAQs please point me in this direction. I have seen a fair amount of threads on related to this topic on this list, but none of them address the equipment that is being used for the collection and analyzing of netflow data. We are looking to collect about Gig of netflow data/hour. We are looking into how we can purge unneeded details, and keep just stats. That said we will atleast want to keeps a weeks worth of data, and backup old data to tape. We most likely will us Solaris as the platform of choice, but are open to others recommendations. What memory and disk storage setups are people using to collect and analyze data set of this size? Any recommendations are welcome. Please respond to me directly. Thanks, -Roy Hockett

2 1

Fix for BSDi TCP Header Issue
by Jonathan A. Zdziarski 03 Nov '98

03 Nov '98

Turns out that BSDi (HTTP and possibly FTP) isn't sending a CR at EOL, but only a LF. That's fine and doesn't break EOL protocol, however the packetshaper isn't pushing the packets through until it gets a CR. The guys at packetshaper are writing a fix for it, but it probably won't hit the release stages for a few weeks. If you're interested in the pre-release fix, and are having the same problems that myself, and other folks are having, contact rick at packeteer support and he can probably get you in on the list. We have about 80% BSDi machines here, so it's a big deal for us including customer platforms. If it's not a big deal for you, I appologize for filling up your nanog box. Thank you, Jonathan A. Zdziarski Sr. Systems Administrator Netrail, inc. 888.NET.RAIL x240

1 0

apology
by Michael Freeman 03 Nov '98

03 Nov '98

I would like to take the time and bandwidth to apologize to everyone who saw my posts about the rootshell hack, my choice of words that I used were wrong and innappropriate. I had a rough day that day dealing with the ssh issues from customers at my work and everything so..no hard feelings anyone. Thanks.

1 0

Xedia vs Packeteer Comparsion
by Michael Gibson 03 Nov '98

03 Nov '98

I wonder if anybody one there has had experience with both xedia and packeteer and would let me know what the strengths and weaknesses was of each Michael Gibson Team Leader, Network Operations - Netcom Canada Telephone: 416-341-5751 Fax: 416-341-5725 magibson(a)netcom.ca

7 9

FW: time for a new list?
by Lawrence A. Deleski 03 Nov '98

03 Nov '98

---------- >From: Brett_Watson(a)enron.net >To: nanog(a)merit.edu >Subject: Re: time for a new list? >Date: Tue, Nov 3, 1998, 5:29 PM > >na... the only thing that will ever work on this list is marshall law. or >better yet, a dictator. > >-brett > > I nominate Brett Watson. :-) Larry Deleski Program Manager Cisco PSO ldeleski(a)cisco.com

1 0

Re: time for a new list?
by Brett_Watson＠enron.net 03 Nov '98

03 Nov '98

na... the only thing that will ever work on this list is marshall law. or better yet, a dictator. -brett Peter Polasek <pete(a)cobra.brass.com> on 11/03/98 07:27:48 AM To: Josh Gluck <josh(a)tricreations.com> cc: nanog(a)merit.edu, Peter Polasek <pete(a)cobra.brass.com> (bcc: Brett Watson/Enron Communications) Subject: Re: time for a new list? I think that all that's needed are some loose guidelines. The sun-managers list, for example, loosely requires that all issues be discussed off-line (i.e. through direct e-mail to the submitter) after an issue has been initially submitted to the list. The submitter is then required to issue a summary of the results after all responses have been received. The only management that is in place (that I know of) is that the group mail-server rejects replies to the list address (which eliminates 95% of off-line discussions from being broadcast to the list). There are some habitual offenders of the rules, but it is largely an effective arrangement. I'm sure there are very many people (myself included) who have an interest in the NANOG discussions but simply can not take the sheer volume of messages. Peter On Tue, 3 Nov 1998, Josh Gluck wrote: > I have been reading this list for a while now, and for the most part, > just sitting back reading posts that interest me. I appreciate the early > heads up on outages, new info on networking technology and the like. > However, I do believe that moderating or "policing" the list would be a good > idea. I hate downloading my mail sometimes, for all the nonsense posts that > have been being sent to NANOG lately. Some of the threads go on forever or > seem to attract many of the following replies: > > { > > From: whoever(a)nowhere.com > To: Nanog(a)merit.edu > Subject: One REALLY LONG THREAD ABOUT NOTHING > > <<< > <<< Some quoted non needed long email from the thread > <<< that everyone has already seen about 20 times > <<< > > Agreed > > :) > > ------------------------------------------------------------ > >A 20 line sig that no one really cares about anyway > ------------------------------------------------------------ > > } > > These kind of replies are not needed. When I see these in my inbox, I > will stop reading the list and just delete everything. Most of us that > receive the list, don't need 3 pages of quoted material from the previous > post. All of the subscribers recieve the threads, I personally dont need it > more than once. Just use the original subject line and throw in a Re: > > As for a subscription requirement, that may be a little difficult as the > "good" information that passes through this list, is interesting to a great > deal of people that may not fit into the subscription group. It wouldn't be > fair to have them be denied from the right to see what is going on, however > a posting credentials requirement might be a good idea, but then again who > would enforce this, and what would the requirements be? > > I like this list a great deal, and use some of the information that I > gather from reading posts. I am willing to add any additional thoughts or > what have you, in an attempt to get NANOG more stream lined. I don't want > people to think that I am standing on a soap box, preaching, especially > because I don't really post to the list, but I am just letting everyone know > I would like to help make the list "a better place" > > > Josh Gluck - josh(a)tricreations.com > Senior Unix System Administrator > Tri Star Web Technical Support > 888-757-4248 > > > *The views expressed are my own and > do not reflect at all on the part of my > employer. > >

1 0

Re: machine specs for Netflow collector & Analyzer
by Brett_Watson＠enron.net 03 Nov '98

03 Nov '98

(this is with my customer login, not the "public" pages) flow collector: http://www.cisco.com/univercd/cc/td/doc/product/rtrmgmt/nfc/nfc_2_0/index.h tm and flow analyer: http://www.cisco.com/univercd/cc/td/doc/product/rtrmgmt/nfa/index.htm you want version 2.0 and the docs under there should have hardware requirements and all the knobs you want for data reduction, etc (ok, maybe not *all* the knobs but it's a nice improvement over v1.0) wow, did i just post something on-topic, and non-drivelous? -brett Roy Hockett <royboy(a)umich.edu> on 11/02/98 10:11:12 PM Please respond to Roy Hockett <royboy(a)umich.edu> To: nanog(a)merit.edu cc: (bcc: Brett Watson/Enron Communications) Subject: machine specs for Netflow collector & Analyzer I know this is not in the spirit of this list, however, I am sure those who have addressed this issue are participants in this list. If you have web resources, or FAQs please point me in this direction. I have seen a fair amount of threads on related to this topic on this list, but none of them address the equipment that is being used for the collection and analyzing of netflow data. We are looking to collect about Gig of netflow data/hour. We are looking into how we can purge unneeded details, and keep just stats. That said we will atleast want to keeps a weeks worth of data, and backup old data to tape. We most likely will us Solaris as the platform of choice, but are open to others recommendations. What memory and disk storage setups are people using to collect and analyze data set of this size? Any recommendations are welcome. Please respond to me directly. Thanks, -Roy Hockett

1 0

BSDi TCP Headers
by Jonathan A. Zdziarski 03 Nov '98

03 Nov '98

After dealing with packeteer for about a week regarding our latest problems, we finally were able to find the root of the problem; several other customers as well as myself all share one common thread; we all are experiencing crawling slow transfers with BSDi (when using the packetshaper). I tried it from a solaris machine on the same network, and transfers immediately picked up. Packeteer did some sniffing, and it appears that BSDi "breaks" the HTTP protocol, and we suspect it's doing the same thing with FTP (as that's where I'm having my problems). It's most likely the way it writes TCP headers rather than the protocol itself (i.e. apache isn't the one at fault). I've got an open issue with packeteer about it, they're trying to come out with a hack for their code to work with BSDi machines, but in the meantime it would be a better idea to solve the problem at the root and find out if there's anything the BSDi folks can do to explore/patch this problem. Is anyone else having similar problems w/BSDi and TCP Transport? Thank you, Jonathan A. Zdziarski Sr. Systems Administrator Netrail, inc. 888.NET.RAIL x240

1 0

Re: Xedia vs Packeteer Comparsion
by Tim Salo 03 Nov '98

03 Nov '98

> Date: Sun, 1 Nov 1998 08:54:04 -0500 (EST) > From: "Jonathan A. Zdziarski" <jonz(a)netrail.net> > Subject: Re: Xedia vs Packeteer Comparsion > [...] > I was told that the xedia boxes only shape outgoing bandwidth, not > incoming. ... > [...] This may be only a matter of what you call "outgoing" bandwidth. Any IP device only has control over how outgoing bandwidth is allocated (where "outgoing" is from the perspective of the device its self, _not_ in the sense of data "outgoing" from a local ISP to a national service provider). An IP device can only control how outgoing bandwidth is allocated by managing its outbound queues, for example deciding which packet to transmit next and which packet to drop next. The device has no control (in the absence of some sort of QoS request or QoS support from the [e.g., ATM] network) over how incoming bandwidth is allocated; all the device can do is receive the next packet that is sent to it on the line. -tjs

1 0