writing histogram display of dropped packets?

writing histogram display of dropped packets?

Post by Gran » Thu, 13 Oct 2005 15:43:22



Hi there,

Given:

grep InpDrop: /var/log/messages \
        | grep -v ICMP \
        | tail -100 \
        | cut -d= -f4- \
        | sed 's/ DF//' \
        | cut -d" " -f9,11 \
        | sed 's/PROTO=//;s/\(UDP\|TCP\) DPT=\([0-9]\+\)/\2\t\1/' \
        | sort -n | uniq -c > net-rad-raw

produced:

$ cat net-rad-raw
     11 135     TCP
      7 139     TCP
     21 445     TCP
      3 1025    TCP
     23 1026    UDP
     17 1027    UDP
      2 1028    UDP
      2 1029    UDP
      1 1030    UDP
      4 1433    TCP
      1 1434    UDP
      1 4297    UDP
      7 4899    TCP

I'd like to see output as:

  135/tcp ***********
  139/tcp *******
  445/tcp *********************
 1025/tcp ***
...
and so on.  

How do that?  More succinct ways to extract data welcome too,
particularly in awk, if it be appropriate.

Thanks,
Grant.

 
 
 

writing histogram display of dropped packets?

Post by Ed Morto » Thu, 13 Oct 2005 22:49:30


 > Hi there,
 >
 > Given:
 >
 > grep InpDrop: /var/log/messages \
 >         | grep -v ICMP \
 >         | tail -100 \
 >         | cut -d= -f4- \
 >         | sed 's/ DF//' \
 >         | cut -d" " -f9,11 \
 >         | sed 's/PROTO=//;s/\(UDP\|TCP\) DPT=\([0-9]\+\)/\2\t\1/' \
 >         | sort -n | uniq -c > net-rad-raw
 >
 > produced:
 >
 > $ cat net-rad-raw
 >      11 135     TCP
 >       7 139     TCP
 >      21 445     TCP
 >       3 1025    TCP
 >      23 1026    UDP
 >      17 1027    UDP
 >       2 1028    UDP
 >       2 1029    UDP
 >       1 1030    UDP
 >       4 1433    TCP
 >       1 1434    UDP
 >       1 4297    UDP
 >       7 4899    TCP
 >
 > I'd like to see output as:
 >
 >   135/tcp ***********
 >   139/tcp *******
 >   445/tcp *********************
 >  1025/tcp ***
 > ...
 > and so on.
 >
 > How do that?  More succinct ways to extract data welcome too,
 > particularly in awk, if it be appropriate.

If you posted your original input and explained what you want to do with
it to produce the above output, you might get a better way of extracting
the data, but I doubt if anyone's going to reverse engineer your script
because, apart from being time-consuming, they'd just reproduce any bugs
in your script.

For the second part:

$ awk '{printf "%10s/%s ",$2,tolower($3);for (i=1;i<=$1;i++)printf
"*";print ""}' net-rad-raw

        135/tcp ***********
        139/tcp *******
        445/tcp *********************
       1025/tcp ***
       1026/udp ***********************
       1027/udp *****************
       1028/udp **
       1029/udp **
       1030/udp *
       1433/tcp ****
       1434/udp *
       4297/udp *
       4899/tcp *******

Regards,

        Ed.

 
 
 

writing histogram display of dropped packets?

Post by Loki Harfag » Fri, 14 Oct 2005 02:25:55


Le Wed, 12 Oct 2005 08:49:30 -0500, Ed Morton a crit?:


>  > Hi there,

>  > Given:

>  > grep InpDrop: /var/log/messages \
>  >         | grep -v ICMP \
>  >         | tail -100 \
>  >         | cut -d= -f4- \
>  >         | sed 's/ DF//' \
>  >         | cut -d" " -f9,11 \
>  >         | sed 's/PROTO=//;s/\(UDP\|TCP\) DPT=\([0-9]\+\)/\2\t\1/' \
>  >         | sort -n | uniq -c > net-rad-raw

>  > produced:

>  > $ cat net-rad-raw
>  >      11 135     TCP
>  >       7 139     TCP
>  >      21 445     TCP
>  >       3 1025    TCP
>  >      23 1026    UDP
>  >      17 1027    UDP
>  >       2 1028    UDP
>  >       2 1029    UDP
>  >       1 1030    UDP
>  >       4 1433    TCP
>  >       1 1434    UDP
>  >       1 4297    UDP
>  >       7 4899    TCP

>  > I'd like to see output as:

>  >   135/tcp ***********
>  >   139/tcp *******
>  >   445/tcp *********************
>  >  1025/tcp ***
>  > ...
>  > and so on.

>  > How do that?  More succinct ways to extract data welcome too,
>  > particularly in awk, if it be appropriate.

> If you posted your original input and explained what you want to do with
> it to produce the above output, you might get a better way of extracting
> the data, but I doubt if anyone's going to reverse engineer your script
> because, apart from being time-consuming, they'd just reproduce any bugs
> in your script.

> For the second part:

> $ awk '{printf "%10s/%s ",$2,tolower($3);for (i=1;i<=$1;i++)printf
> "*";print ""}' net-rad-raw

>         135/tcp ***********
>         139/tcp *******
>         445/tcp *********************
>        1025/tcp ***
>        1026/udp ***********************
>        1027/udp *****************
>        1028/udp **
>        1029/udp **
>        1030/udp *
>        1433/tcp ****
>        1434/udp *
>        4297/udp *
>        4899/tcp *******

 Neat, I'd try this slight simplification (one only loop on stars)
$ awk 'BEGIN{for(i=1;i<101;i++)pups=pups"*";}
    {printf "%10s/%s ",$2,tolower($3);print substr(pups,1,$1)}'

  Now, it's failure prone if the numbers aren't levelled to pp100,
hope we'll soon know more about the original data :-)

 
 
 

writing histogram display of dropped packets?

Post by Ed Morto » Fri, 14 Oct 2005 02:36:47



> Le Wed, 12 Oct 2005 08:49:30 -0500, Ed Morton a crit :
<snip>
>>$ awk '{printf "%10s/%s ",$2,tolower($3);for (i=1;i<=$1;i++)printf
>>"*";print ""}' net-rad-raw
<snip>

>  Neat, I'd try this slight simplification (one only loop on stars)
> $ awk 'BEGIN{for(i=1;i<101;i++)pups=pups"*";}
>     {printf "%10s/%s ",$2,tolower($3);print substr(pups,1,$1)}'

>   Now, it's failure prone if the numbers aren't levelled to pp100,
> hope we'll soon know more about the original data :-)

Don't know if it's any better, but you could also do it this way and
avoid a loop completely:

$ awk '{s=sprintf("%"$1"s","");gsub(" ","*",s);printf "%10s/%s
%s\n",$2,tolower($3),s}' net-rad-raw
        135/tcp ***********
        139/tcp *******
        445/tcp *********************
       1025/tcp ***
       1026/udp ***********************
       1027/udp *****************
       1028/udp **
       1029/udp **
       1030/udp *
       1433/tcp ****
       1434/udp *
       4297/udp *
       4899/tcp *******

Regards,

        Ed.

 
 
 

writing histogram display of dropped packets?

Post by William Jame » Fri, 14 Oct 2005 02:51:29



> Hi there,

> Given:

> grep InpDrop: /var/log/messages \
>         | grep -v ICMP \
>         | tail -100 \
>         | cut -d= -f4- \
>         | sed 's/ DF//' \
>         | cut -d" " -f9,11 \
>         | sed 's/PROTO=//;s/\(UDP\|TCP\) DPT=\([0-9]\+\)/\2\t\1/' \
>         | sort -n | uniq -c > net-rad-raw

> produced:

> $ cat net-rad-raw
>      11 135     TCP
>       7 139     TCP
>      21 445     TCP
>       3 1025    TCP
>      23 1026    UDP
>      17 1027    UDP
>       2 1028    UDP
>       2 1029    UDP
>       1 1030    UDP
>       4 1433    TCP
>       1 1434    UDP
>       1 4297    UDP
>       7 4899    TCP

> I'd like to see output as:

>   135/tcp ***********
>   139/tcp *******
>   445/tcp *********************
>  1025/tcp ***
> ...
> and so on.

> How do that?  More succinct ways to extract data welcome too,
> particularly in awk, if it be appropriate.

> Thanks,
> Grant.

To produce the histogram:

ruby -nae'printf "%7d/%s %s\n", $F[1], $F[2], "*" * $F[0].to_i'

 
 
 

writing histogram display of dropped packets?

Post by Loki Harfag » Fri, 14 Oct 2005 03:14:34


Le Wed, 12 Oct 2005 12:36:47 -0500, Ed Morton a crit?:


>> Le Wed, 12 Oct 2005 08:49:30 -0500, Ed Morton a crit :
> <snip>
>>>$ awk '{printf "%10s/%s ",$2,tolower($3);for (i=1;i<=$1;i++)printf
>>>"*";print ""}' net-rad-raw
> <snip>

>>  Neat, I'd try this slight simplification (one only loop on stars)
>> $ awk 'BEGIN{for(i=1;i<101;i++)pups=pups"*";}
>>     {printf "%10s/%s ",$2,tolower($3);print substr(pups,1,$1)}'

>>   Now, it's failure prone if the numbers aren't levelled to pp100,
>> hope we'll soon know more about the original data :-)

> Don't know if it's any better, but you could also do it this way and
> avoid a loop completely:

> $ awk '{s=sprintf("%"$1"s","");gsub(" ","*",s);printf "%10s/%s
> %s\n",$2,tolower($3),s}' net-rad-raw
>         135/tcp ***********
>         139/tcp *******
>         445/tcp *********************
>        1025/tcp ***
>        1026/udp ***********************
>        1027/udp *****************
>        1028/udp **
>        1029/udp **
>        1030/udp *
>        1433/tcp ****
>        1434/udp *
>        4297/udp *
>        4899/tcp *******

  That's right, I thought about this kind of solution too, but decided,
maybe wrongly, that there was indeed a loop undercover inside the 'gsub' ;-)
  Not that I think it'd be a real stress to the system as the input file
is certainly less than 2*65535 lines :D)
 
 
 

writing histogram display of dropped packets?

Post by John W. Krah » Fri, 14 Oct 2005 03:20:28



> Given:

> grep InpDrop: /var/log/messages \
>         | grep -v ICMP \
>         | tail -100 \
>         | cut -d= -f4- \
>         | sed 's/ DF//' \
>         | cut -d" " -f9,11 \
>         | sed 's/PROTO=//;s/\(UDP\|TCP\) DPT=\([0-9]\+\)/\2\t\1/' \
>         | sort -n | uniq -c > net-rad-raw

> produced:

> $ cat net-rad-raw
>      11 135     TCP
>       7 139     TCP
>      21 445     TCP
>       3 1025    TCP
>      23 1026    UDP
>      17 1027    UDP
>       2 1028    UDP
>       2 1029    UDP
>       1 1030    UDP
>       4 1433    TCP
>       1 1434    UDP
>       1 4297    UDP
>       7 4899    TCP

> I'd like to see output as:

>   135/tcp ***********
>   139/tcp *******
>   445/tcp *********************
>  1025/tcp ***
> ...
> and so on.  

> How do that?  More succinct ways to extract data welcome too,
> particularly in awk, if it be appropriate.

This seems to work for me.  :-)

$ perl -ne'
BEGIN { $#x = 99 }

END {

    printf "%10s %s\n", $_, "*" x $x{$_} for sort { $a <=> $b } keys %x;
    }
' /var/log/messages
    80/tcp *******
   119/tcp *
   135/tcp **
   445/tcp *****
  1025/udp ****
  1026/udp *******************************
  1027/udp ********************
  1028/udp ******
  1029/udp ******
  1434/udp **
  3112/tcp ********
  5353/udp **
 10000/tcp **
 11768/tcp ****

John
--
use Perl;
program
fulfillment

 
 
 

writing histogram display of dropped packets?

Post by John W. Krah » Fri, 14 Oct 2005 03:26:43



> Given:

> grep InpDrop: /var/log/messages \
>         | grep -v ICMP \
>         | tail -100 \
>         | cut -d= -f4- \
>         | sed 's/ DF//' \
>         | cut -d" " -f9,11 \
>         | sed 's/PROTO=//;s/\(UDP\|TCP\) DPT=\([0-9]\+\)/\2\t\1/' \
>         | sort -n | uniq -c > net-rad-raw

> produced:

> $ cat net-rad-raw
>      11 135     TCP
>       7 139     TCP
>      21 445     TCP
>       3 1025    TCP
>      23 1026    UDP
>      17 1027    UDP
>       2 1028    UDP
>       2 1029    UDP
>       1 1030    UDP
>       4 1433    TCP
>       1 1434    UDP
>       1 4297    UDP
>       7 4899    TCP

> I'd like to see output as:

>   135/tcp ***********
>   139/tcp *******
>   445/tcp *********************
>  1025/tcp ***
> ...
> and so on.  

> How do that?  More succinct ways to extract data welcome too,
> particularly in awk, if it be appropriate.

This seems to work for me.  :-)

# perl -ne'
BEGIN { $#x = 99 }
next unless /InpDrop:/;

END {

    printf "%10s %s\n", $_, "*" x $x{$_} for sort { $a <=> $b } keys %x;
    }
' /var/log/messages
    80/tcp *******
   119/tcp *
   135/tcp **
   445/tcp *****
  1025/udp ****
  1026/udp *******************************
  1027/udp ********************
  1028/udp ******
  1029/udp ******
  1434/udp **
  3112/tcp ********
  5353/udp **
 10000/tcp **
 11768/tcp ****

John
--
use Perl;
program
fulfillment

 
 
 

writing histogram display of dropped packets?

Post by Chris F.A. Johnso » Fri, 14 Oct 2005 04:08:18



> Hi there,

> Given:

> grep InpDrop: /var/log/messages \
>         | grep -v ICMP \
>         | tail -100 \
>         | cut -d= -f4- \
>         | sed 's/ DF//' \
>         | cut -d" " -f9,11 \
>         | sed 's/PROTO=//;s/\(UDP\|TCP\) DPT=\([0-9]\+\)/\2\t\1/' \
>         | sort -n | uniq -c > net-rad-raw

> produced:

> $ cat net-rad-raw
>      11 135     TCP
>       7 139     TCP
>      21 445     TCP
>       3 1025    TCP
>      23 1026    UDP
>      17 1027    UDP
>       2 1028    UDP
>       2 1029    UDP
>       1 1030    UDP
>       4 1433    TCP
>       1 1434    UDP
>       1 4297    UDP
>       7 4899    TCP

> I'd like to see output as:

>   135/tcp ***********
>   139/tcp *******
>   445/tcp *********************
>  1025/tcp ***
> ...
> and so on.  

awk 'BEGIN { ## Adjust number of stars to taste
     stars = "**************************************************************"
           }
           { printf "%5d/%-3.3s %s\n", $2, $3, substr(stars,1,$1) }'

    That doesn't convert TCP or UDP to lower case, but you can do that
    by piping the data through 'tr '[A-Z]' '[a-z]', or with a POSIX
    awk, by changing $3 to tolower($3).

Quote:> How do that?  More succinct ways to extract data welcome too,
> particularly in awk, if it be appropriate.

    To recommend better data extraction methods, I'd have to see what
    'grep InpDrop: /var/log/messages' looks like.

--
    Chris F.A. Johnson                     <http://cfaj.freeshell.org>
    ==================================================================
    Shell Scripting Recipes: A Problem-Solution Approach, 2005, Apress
    <http://www.torfree.net/~chris/books/cfaj/ssr.html>

 
 
 

writing histogram display of dropped packets?

Post by Gran » Fri, 14 Oct 2005 06:40:29



>If you posted your original input and explained what you want to do with
>it to produce the above output, you might get a better way of extracting
>the data, but I doubt if anyone's going to reverse engineer your script
>because, apart from being time-consuming, they'd just reproduce any bugs
>in your script.

Okay, fair enough :)

data source, /var/log/messages single line sample:
Oct 13 07:16:34 deltree kernel: InpDrop: IN=ppp0 OUT= MAC= SRC=123.123.123.123 DST=123.123.123.123 LEN=48 TOS=0x00 PREC=0x20 TTL=126 ID=62514 DF PROTO=TCP SPT=4301 DPT=139 WINDOW=65535 RES=0x00 SYN URGP=0

The " DF" field is not always present in log entries, thus use of '=' as
delimiter.

task:
sample most recent 100 PROTO=TCP or PROTO=UDP and display as histogram
indicator of network 'background radiation'.  Perhaps present as top-like
current, last minute, last 1/4 hour, side by side on a text web page :o)

Quote:

>For the second part:

>$ awk '{printf "%10s/%s ",$2,tolower($3);for (i=1;i<=$1;i++)printf
>"*";print ""}' net-rad-raw

>        135/tcp ***********
>        139/tcp *******

Thank you Ed, something like that I can build on.

Grant.

 
 
 

writing histogram display of dropped packets?

Post by Gran » Fri, 14 Oct 2005 06:51:16



> Neat, I'd try this slight simplification (one only loop on stars)
>$ awk 'BEGIN{for(i=1;i<101;i++)pups=pups"*";}
>    {printf "%10s/%s ",$2,tolower($3);print substr(pups,1,$1)}'

>  Now, it's failure prone if the numbers aren't levelled to pp100,
>hope we'll soon know more about the original data :-)

There's 0..100 samples due to log-rotate :(

Grant.

 
 
 

writing histogram display of dropped packets?

Post by Gran » Fri, 14 Oct 2005 07:05:47



...

Quote:>awk 'BEGIN { ## Adjust number of stars to taste
>     stars = "**************************************************************"
>           }
>           { printf "%5d/%-3.3s %s\n", $2, $3, substr(stars,1,$1) }'

This is more what I was looking for, ideas to avoid looping, didn't
think of string-chopping like that :o)

Quote:>    That doesn't convert TCP or UDP to lower case, but you can do that
>    by piping the data through 'tr '[A-Z]' '[a-z]', or with a POSIX
>    awk, by changing $3 to tolower($3).

Sure.

Quote:>    To recommend better data extraction methods, I'd have to see what
>    'grep InpDrop: /var/log/messages' looks like.

Oct 13 07:16:34 deltree kernel: InpDrop: IN=ppp0 OUT= MAC= SRC=123.123.123.123 DST=123.123.123.123 LEN=48 TOS=0x00 PREC=0x20 TTL=126 ID=62514 DF PROTO=TCP SPT=4301 DPT=139 WINDOW=65535 RES=0x00 SYN URGP=0

Thanks,
Grant.

 
 
 

writing histogram display of dropped packets?

Post by Ed Morto » Fri, 14 Oct 2005 13:32:39




>>If you posted your original input and explained what you want to do with
>>it to produce the above output, you might get a better way of extracting
>>the data, but I doubt if anyone's going to reverse engineer your script
>>because, apart from being time-consuming, they'd just reproduce any bugs
>>in your script.

> Okay, fair enough :)

> data source, /var/log/messages single line sample:
> Oct 13 07:16:34 deltree kernel: InpDrop: IN=ppp0 OUT= MAC= SRC=123.123.123.123 DST=123.123.123.123 LEN=48 TOS=0x00 PREC=0x20 TTL=126 ID=62514 DF PROTO=TCP SPT=4301 DPT=139 WINDOW=65535 RES=0x00 SYN URGP=0

> The " DF" field is not always present in log entries, thus use of '=' as
> delimiter.

> task:
> sample most recent 100 PROTO=TCP or PROTO=UDP and display as histogram
> indicator of network 'background radiation'.  Perhaps present as top-like
> current, last minute, last 1/4 hour, side by side on a text web page :o)

Well, given this was your original script:

grep InpDrop: /var/log/messages \
         | grep -v ICMP \
         | tail -100 \
         | cut -d= -f4- \
         | sed 's/ DF//' \
         | cut -d" " -f9,11 \
         | sed 's/PROTO=//;s/\(UDP\|TCP\) DPT=\([0-9]\+\)/\2\t\1/' \
         | sort -n | uniq -c > net-rad-raw

it appears that you could replace everything before the "| sort" with this:

awk -F"[ =]" '/ICMP/{next}
/InpDrop:/{n=NR%100;sub(/^.*PROTO=/,"");key[n]=$1;dpt[n]=$5}
END{for (i in dpt) print dpt[i],proto[i]}' /var/log/messages

If you're using gawk, you can do the sort within "awk" (using asort())
but that hardly seems worthwhile. If you don't care about the sort, then
you can finish it off entirely in awk:

awk -F"[ =]" '/ICMP/{next}
/InpDrop:/{sub(/^.*PROTO=/,"");key[NR%100]=$5"/"$1}
END{for (i in key)
        cnt[key[i]]++
     for (i in cnt) {
        printf "%10s ",tolower(i)
        for (j=1;j<=cnt[i];j++)
           printf "*"
        print ""
     }

Quote:}' /var/log/messages

Regards,

        Ed.

 
 
 

writing histogram display of dropped packets?

Post by Loki Harfag » Fri, 14 Oct 2005 16:11:18


Le Thu, 13 Oct 2005 08:05:47 +1000, Grant a crit?:


> ...
>>awk 'BEGIN { ## Adjust number of stars to taste
>>     stars = "**************************************************************"
>>           }
>>           { printf "%5d/%-3.3s %s\n", $2, $3, substr(stars,1,$1) }'

> This is more what I was looking for, ideas to avoid looping, didn't
> think of string-chopping like that :o)

 Hem, then you didn't really read my proposal, based
on Ed. solution without the loops :D)

(
$ awk 'BEGIN{for(i=1;i<101;i++)pups=pups"*";}
    {printf "%10s/%s ",$2,tolower($3);print substr(pups,1,$1)}'
)

 The two differences are mainly :
        .I used one loop in BEGIN o replace typeg 100 stars (so laaazy ;-)
        .I used 2 'print[f]', Chris F.A. syntax on a one only
        properly formatted printf is really cleaner and better :D)

 
 
 

writing histogram display of dropped packets?

Post by Chris F.A. Johnso » Fri, 14 Oct 2005 16:41:30



> Le Thu, 13 Oct 2005 08:05:47 +1000, Grant a crit?:


>> ...
>>>awk 'BEGIN { ## Adjust number of stars to taste
>>>     stars = "**************************************************************"
>>>           }
>>>           { printf "%5d/%-3.3s %s\n", $2, $3, substr(stars,1,$1) }'

>> This is more what I was looking for, ideas to avoid looping, didn't
>> think of string-chopping like that :o)

>  Hem, then you didn't really read my proposal, based
> on Ed. solution without the loops :D)

> (
> $ awk 'BEGIN{for(i=1;i<101;i++)pups=pups"*";}
>     {printf "%10s/%s ",$2,tolower($3);print substr(pups,1,$1)}'
> )

>  The two differences are mainly :
>    .I used one loop in BEGIN o replace typeg 100 stars (so laaazy ;-)

    You only need a few iterations, not 101. ;)

BEGIN {for(;n++<7;)pups=pups pups "*"}

Quote:>    .I used 2 'print[f]', Chris F.A. syntax on a one only
>    properly formatted printf is really cleaner and better :D)

--
    Chris F.A. Johnson                     <http://cfaj.freeshell.org>
    ==================================================================
    Shell Scripting Recipes: A Problem-Solution Approach, 2005, Apress
    <http://www.torfree.net/~chris/books/cfaj/ssr.html>
 
 
 

1. Dropped packets, bogus packets and errors with SMC Elite 16 Plus

Hi,

All my systems are using the SMC Elite 16 Plus as network board, since I'm
using samba (which means more load) I got packet errors.

When I do a file copy from an win95 machine to the linux box running samba,
I got approx 1 bogus packet in 5 minutes and the dropped packets and errors
count is also slowly increasing.

I tried this for 3 machines running linux, they all have the same problem.
Novell Netware has no problems with the SMC Elite 16 Plus.
I also checked the wiring, seems to be ok..

eth0      Link encap:10Mbps Ethernet  HWaddr 00:00:C0:F5:7D:62
          inet addr:10.0.0.3  Bcast:10.0.0.255  Mask:255.255.255.0
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:98 errors:3 dropped:155 overruns:0
          TX packets:96 errors:0 dropped:0 overruns:0
          Interrupt:10 Base address:0x250 Memory:d0000-d4000

eth0: WD80x3 at 0x240,  00 00 C0 F5 7D 62 WD8013, IRQ 10,
shared memory at 0xd0000-0xd3fff.

Chipset 8013 EPC

Bogus error messages:
eth0: bogus packet: status=0x0 nxpg=0x2e size=1518
eth0: bogus packet: status=0x0 nxpg=0x19 size=1518

Kernels tried: 20.0.28/30/31

The system does not crash and the network speed is ok.

Thanks in advance,

Gerhard.
--
Gerhard Ahuis          JO32EO               It's good to be independent

Unsolicited advertisements subject to $1000 consulting fee.

2. __ x86 install: stuck at "Configuring the /devices directory" __

3. minixfs patch

4. Me To Linux: Stop Dropping Packets!

5. Super IDE problem (2 hd => 1 logical)

6. Power Mac G3 dropping packets

7. Strange sound

8. flood ping dropping packets localhost

9. Dropped RX Packets on Interface ppp0

10. Router dropping packets from eth interface to ppp interface

11. analysing dropped packet

12. Wireless connection between router and bridge is dropping packets in spurts