FAQ Search Today's Posts Mark Forums Read
» Video Reviews

» Linux Archive

Linux-archive is a website aiming to archive linux email lists and to make them easily accessible for linux users/developers.


» Sponsor

» Partners

» Sponsor

Go Back   Linux Archive > Ubuntu > Kubuntu User

 
 
LinkBack Thread Tools
 
Old 12-28-2010, 02:23 PM
"Boggess Rod"
 
Default Network problems

>-----Original Message-----
>From: Bill Stanley [mailto:bstanle@wowway.com]
>Sent: Monday, December 27, 2010 10:31 PM
>To: Boggess Rod
>Subject: Re: Network problems
>
>On 12/27/2010 02:33 PM, Boggess Rod wrote:
>>> By the way, it has been suggested that I do a reinstall on the
affected
>>> computer... I am getting to that point but not just yet. You see, I
>>> learn the most when confronting a problem and I am certainly
learning a
>>> lot now. When my frustration reaches a certain point, I will
>> reinstall.
>>>
>> Now that I'm caught up on the emails, a quick recap:
>>
>> 1) This is primarily a learning experience, as opposed to a
production
>> setup that you have to have running no later than yesterday.
>
>WS=> This is correct. I did say that this was a home network. A bit
of
>background about my situation. I am presently retired (disability) but
>I do have an advanced degree in computer science. I had to quit my PhD
>dissertation (I was just about to start) because of my illness which in
>no was affects my mental abilities. I am presently working with my
PhD
>committee chairman (but not for course credit). As such I have no
real
>deadlines. The problem is that computer networking is far from my
>specialty. I have had the obligatory network courses but I have just
>scratched the surface on network.s (My specialty is testing, verifying
>and debugging software.)
>
>> 2) You have (at present) two computers running Linux.
>
>I actually have a third computer that I just today installed Unbuntu
>10.10 on. It is very ancient and I am suprised that I was actually
>gotten Unbuntu to work on. As a measure of what I mean by ancient.
The
>third computer has a Pentium II processor running at 300 Mz. It has
384
>Mb of RAM and the linux partition is Ext4 and 6.5 Gbyte in size. It is
>running very SLOW but it is running. I have installed the WireShark
>network analyzer on it and I intend to use it to record the network
>communications of the other two computers. It should be able to record
>their communications. I might try to see if can share files with it
>but I am not holding my breath.

You might try Xubuntu instead of Ubuntu since it's more resource
friendly. Xubuntu uses the Xfce x-manager which has lower demands than
GNOME, but it's similar enough that you won't feel lost. Using Wireshark
is fun for a learning experience, but I'm not really sure what
troubleshooting value you'll find in it here. You should start off with
something that works first, so you can see how the network performs
normally. Then you can more readily spot changes.

Note that if you're using a switch in full-duplex that non-broadcast
traffic (most of what you want to see) is filtered out by the switch. In
full duplex, once the DNS resolves to IP and ARP resolves to MAC
address, all direct communications should take place between these two
ports on the switch. The third port, the one where your wireshark is
running, will not be "bothered" with this traffic. To change this, you
can either use a really cheap (unintelligent) router (hard to find); you
can (usually) configure your switch to capture all traffic on one port
and connect the wireshark computer to that port; or you can run
wireshark from one of the two machines participating in the
conversation. (I have always done the last part, so I can't help much
with the setting up promiscuous ports on a router or switch. Simplex
mode on the switch might also produce this behavior on all ports, but I
don't know.)

There are web sites and lists dedicated to wireshark, and you'll find
them very friendly. I have to mention, though, that wireshark also runs
on Windows, so you needn't have installed linux at all on the other box
for this. The windows version is harder to run the Ethernet card in
promiscuous mode, requiring an update to a system DLL, but it also has
features in the Windows version that are not available in the linux
distribution.

>
>> 3) Ping works for name and IP for one, but only IP for the other. For
>> the one that does work, IP by host name is more than a hundred times
>> slower and does not resolve to the correct, desired IP.
>
>Your question about ping. I can successfully ping the ancient computer
>and vice versa when I specify their ip numbers but not when I specify
>their host-names. Maybe that is because I am not running name server
>software? At least this test shows that it is not the router or the
>network per sat. As such, it has to be something with the second
>computer. If necessary, I can transfer files via CD rOM but this is
>both costly and time consuming. Most of my important files are still
on
>the second computer.

Converting a hostname to an IP address is performed by the Domain Name
Server (DNS). So, yes, that's your first mystery solved. I'm guessing
you have a DNS specified in the configuration, but that the DNS server
is external. However, if you're running DHCP, sometimes the DNS is built
into the DHCP server for those addresses it assigns. If you're using
fixed IP addressing, this doesn't apply, but if not, you might try to
determine the IP address of your Domain Host Configuration Protocol
(DHCP) server and add that IP address to the list of DNS servers on both
machines, then try pinging by host name again. In this case, you'll want
the local DNS to be first in the list. Thus, if there's a name
resolution conflict, it'll get your local machine first; but this is
also quicker on average (on the assumption that most traffic will be
local).

Agreed, sneakernet sucks. If it comes to that, though (and it
shouldn't), don't you have a usb stick? Quicker than CD ROM. Still, if
the fix lacks urgency, you'll eventually figure it out. Worst case, you
should be able to create -- I believe it's called a crossover cable --
to connect the two computers directly. (For that matter, it might be
quicker to open the case and pull the drive to do a direct copy.)

>
>> 4) You have double-checked cabling, but not the router.
>
>My successful ping test with the ancient computer indicates that it
>isn't a router or cabling problem. I have not yet tried to share fles
>between these computers.
>
>> 5) You're adding a third (windows) computer.
>
>The third computer can dual boot to Win 2000. I have today installed
>Unbuntu 10.10 on it and am presently using it as a Linux computer. By
>the way, I can dual boot to windows on all three computers.
>
>> Is that about right? Or am I off on that third or fourth bit? Are you
>> using fixed IP or are you running DHCP in the router or one of these
>> machines? Can we assume that the firewall has no filtering enabled,
>> neither ports nor SNMP replies?
>
>I am running DCHP in the router. From my discussion with you all, I
>gather that I do not have any rules. ( I might be wrong although I
have
>been playing around with the firewall on computer two. Where can I
look
>to see what, if any, rules are written? As for Samba replies, that is
>what the WireShark network analyzer will do.
>
>It is getting late so I will continue this battle tomorrow. Right now,
>one has to see the mess of cables around my desk with 3 computers
>running. It's quite confusing to be sure I am using the right keyboard
>and mouse. Anything you might say might set me on the right track to
>resolving this problem. My saga will continue tomorrow. Thanks again
>for any help you might offer,

I'm pretty sure of a few things: 1) you don't have SAMBA server running.
If you did, the port would have been listed as open in your scan. You
should get samba server running on all of the machines. (Strictly
speaking, this might be overkill, but we can return to this issue
later.) 2) the firewall likely isn't the obstruction.

You mentioned above that you're using a router -- is the router
assigning IP addresses? Can you give a better picture of the network
layout?

--
ubuntu-users mailing list
ubuntu-users@lists.ubuntu.com
Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-users
 
Old 03-01-2012, 03:15 PM
 
Default network problems

We just moved a user, who deals with a *lot* of data, to a new server,
since his old NFS home directory was on a disk that had started showing
problems.

Now, i/o is about six times slower, my manager reports.

After a fair bit of googling, I started looking at tc and ip, and found
the following: from ip address show, first, on the old home directory
server,
eth0: <BROADCAST,MULTICAST> mtu 1500 qdisc pfifo_fast qlen 1000
then, on the new:
eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP qlen 1000

I've tried tc qdisc add dev eth0 mq pfifo_fast
and
tc qdisc add dev eth0 pfifo_fast

and neither works. I've googled, and seen something about you can't set
something with tc, because pfifo_fast is the hardwired default.

Anyone know how I can reset this, or is there a package I can reinstall
that would do that?

mark


_______________________________________________
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos
 
Old 03-01-2012, 03:39 PM
John Doe
 
Default network problems

From: "m.roth@5-cent.us" <m.roth@5-cent.us>

> I've tried tc qdisc add dev eth0 mq pfifo_fast
> and tc qdisc add dev eth0* pfifo_fast
> and neither works. I've googled, and seen something about you can't set
> something with tc, because pfifo_fast is the hardwired default.

Not an expert, but this seems to add 'mq' succesfully on my unused eth1:
eth1: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN qlen 1000
# tc qdisc add dev eth1 root handle 1: mq
eth1: <BROADCAST,MULTICAST> mtu 1500 qdisc mq state DOWN qlen 1000

But it fails for 'pfifo_fast':
# tc qdisc add dev eth1 root handle 1: pfifo_fast
qdisc 'pfifo_fast' does not support option parsing

Also, maybe you need to delete before adding?

JD
_______________________________________________
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos
 
Old 03-01-2012, 03:45 PM
John Doe
 
Default network problems

From: John Doe <jdmls@yahoo.com>

> But it fails for 'pfifo_fast':

But it works for 'pfifo'...

JD
_______________________________________________
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos
 
Old 03-01-2012, 03:53 PM
 
Default network problems

John Doe wrote:
> From: "m.roth@5-cent.us" <m.roth@5-cent.us>
>
>> I've tried tc qdisc add dev eth0 mq pfifo_fast
>> and tc qdisc add dev eth0* pfifo_fast
>> and neither works. I've googled, and seen something about you can't set
>> something with tc, because pfifo_fast is the hardwired default.
>
> Not an expert, but this seems to add 'mq' succesfully on my unused eth1:
> eth1: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN qlen 1000
> # tc qdisc add dev eth1 root handle 1: mq
> eth1: <BROADCAST,MULTICAST> mtu 1500 qdisc mq state DOWN qlen 1000
>
> But it fails for 'pfifo_fast':
> # tc qdisc add dev eth1 root handle 1: pfifo_fast
> qdisc 'pfifo_fast' does not support option parsing
>
> Also, maybe you need to delete before adding?

Yeah: after I tried to add it, and it gave me the "does not support
options parsing", I began trying to del the mq, but I can't seem to find
the Magical Syntax (tm) that let's me get rid of that.

mark

_______________________________________________
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos
 
Old 03-01-2012, 09:29 PM
Nataraj
 
Default network problems

On 03/01/2012 08:15 AM, m.roth@5-cent.us wrote:
> We just moved a user, who deals with a *lot* of data, to a new server,
> since his old NFS home directory was on a disk that had started showing
> problems.
>
> Now, i/o is about six times slower, my manager reports.
>
> After a fair bit of googling, I started looking at tc and ip, and found
> the following: from ip address show, first, on the old home directory
> server,
> eth0: <BROADCAST,MULTICAST> mtu 1500 qdisc pfifo_fast qlen 1000
> then, on the new:
> eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP qlen 1000
>
> I've tried tc qdisc add dev eth0 mq pfifo_fast
> and
> tc qdisc add dev eth0 pfifo_fast
>
> and neither works. I've googled, and seen something about you can't set
> something with tc, because pfifo_fast is the hardwired default.
>
> Anyone know how I can reset this, or is there a package I can reinstall
> that would do that?
>
> mark
>
>
> _______________________________________________
> CentOS mailing list
> CentOS@centos.org
> http://lists.centos.org/mailman/listinfo/centos
Are you sure your performance problems are network related? Have you
measured performance of the filesystems on the two servers? You may
have a data alignment problem with your disk partitions.

Nataraj

_______________________________________________
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos
 
Old 03-01-2012, 09:31 PM
 
Default network problems

Nataraj wrote:
> On 03/01/2012 08:15 AM, m.roth@5-cent.us wrote:
>> We just moved a user, who deals with a *lot* of data, to a new server,
>> since his old NFS home directory was on a disk that had started showing
>> problems.
>>
>> Now, i/o is about six times slower, my manager reports.
>>
>> After a fair bit of googling, I started looking at tc and ip, and found
>> the following: from ip address show, first, on the old home directory
>> server,
>> eth0: <BROADCAST,MULTICAST> mtu 1500 qdisc pfifo_fast qlen 1000
>> then, on the new:
>> eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP qlen
>> 1000
<snip>
> Are you sure your performance problems are network related? Have you
> measured performance of the filesystems on the two servers? You may
> have a data alignment problem with your disk partitions.

He did. It's not disk-related.

mark

_______________________________________________
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos
 
Old 03-02-2012, 09:17 AM
John Doe
 
Default network problems

From: "m.roth@5-cent.us" <m.roth@5-cent.us>

> Yeah: after I tried to add it, and it gave me the "does not support
> options parsing", I began trying to del the mq, but I can't seem to
> find the Magical Syntax (tm) that let's me get rid of that.

This worked for me...
# tc qdisc del dev eth1 root handle 1

JD
_______________________________________________
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos
 
Old 03-02-2012, 01:48 PM
 
Default network problems

(Sorry 'bout the screwed-up headers - forwarded to my work email from
home, then forwarded back to this account just now....)

---------------------------- Original Message ----------------------------
From: John Doe <jdmls@yahoo.com>

References:
<76baacf79f928a056c17d63f02d53ae7.squirrel@mail. 5-cent.us>
<1330619990.16821.YahooMailNeo@web114705.mail.gq1. yahoo.com>
<86feb08093663823b4190d6fae1e3a3b.squirrel@mail. 5-cent.us>

From: "m.roth@5-cent.us" <m.roth@5-cent.us>

> Yeah: after I tried to add it, and it gave me the "does not support
> options parsing", I began trying to del the mq, but I can't seem to
> find the Magical Syntax (tm) that let's me get rid of that.

This worked for me...
# tc qdisc del dev eth1 root handle 1

That works... but won't get rid of mq. Right now, I'm trying pfifo, unless
someone thinks that PRIO might be better.

The real question is why *some* of my 6.2 boxen have pfifo_fast, and
others have mq. Does anyone know where this is set?

mark

_______________________________________________
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos
 
Old 03-02-2012, 08:19 PM
 
Default network problems

Nataraj wrote:
> On 03/01/2012 08:15 AM, m.roth@5-cent.us wrote:
>> We just moved a user, who deals with a *lot* of data, to a new server,
since his old NFS home directory was on a disk that had started showing
problems.
>>
>> Now, i/o is about six times slower, my manager reports.
<snip>
Ok, more info: at this point, it seems obvious that it's an NFSv4 problem,
not a network problem: I've created /scratch/foo, exported it, and mounted
it on the same machine at /mnt/foo. I then cd to that directory, and
unpack a large file which is located in a user's home directory which is
NFSv4 mounted. In a native directory, takes about a minute; when I cd to
/mnt/foo and do it, it takes about 6.5 min. I've tried setting wsize and
rsize to 1M; I've tried noatime, which someone in a post said helped, and
nothing I do makes more than a second or two's difference.

Clues for the poor?

mark

_______________________________________________
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos
 

Thread Tools




All times are GMT. The time now is 07:13 PM.

VBulletin, Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
Content Relevant URLs by vBSEO ©2007, Crawlability, Inc.
Copyright 2007 - 2008, www.linux-archive.org