Redhat linux commands

autoscan freshmeat

drbd freshmeat network raid 1

snap for easy video with otherwise unsupported hardware on linux

mii-tool is the old way.  It doesn't work on gigabit networks

ethtool works on 10, 100 and 1000 BaseT

To change the MTU on an RHEL 3 system:
vi /etc/sysconfig/network-scripts/ifcfg-eth*
And add a line like "MTU=9000".  It should take effect on your next reboot.


From Duncan:

I've set up a system that allows http://www.cgsecurity.org/index.html?testdisk.html
DCS team members to 
add and remove hosts to which the RHEL O/S images
will be exported to without having to log in to rhn
or be root on that system.

You can edit the file on autoinst.nacs.uci.edu--it's
called /auto_install/rhel/export/rhn-hosts.

That directory is automounted from rhn.nacs and a cron
job runs every 15 minutes starting on the hour to push
out the changes.

Registration is a pain because it seems impossible to get
around registering using their GUI but if you select 

   'Tell me why I should register'

then

   'Can not register at this time'


You'll be able to continue through the post-install
and then register with the key.


cdrecord -scanbus said 1,0,0, but this was what
actually worked on tesuji, with kernel 2.6.7:

cdrecord -v -dev ATAPI:0,0,0 securepoint-4.5p3.iso

Sometimes this works better (linus' 2.6.7 kernel on FC2, tesuji) :
cdrecord -v -dev=ATA:1,0,0 bootcd.iso

Try the "-eject" option.  :)

mkisofs -o cd.iso -hfs -J -r -graft-points pixtosend

tracepath is nice for examining pmtu, but apparently it sometimes lies
- because I can see larger frames than tracepath reports going between
the hosts in question, through tethereal.

screen -r -x should share a screen session

nfsstat -o rpc show rpc stats

-bash-2.05b# cat /proc/net/rpc/nfsd
rc 1386 1037343 420483406
fh 979922 420473898 0 9497 327249
io 405027549 4193151430
th 12 7859408 122106.880 46341.950 22599.080 36481.660 32597.670 26455.890
9471.830 10035.580 1302.590 13274.840
ra 24 405758114 3293 859 131 55 84 55 17 11 8 82534
net 421522940 421522940 0 0
rpc 421522135 805 805 0 0
proc2 18 811 695 0 0 25344 12 0 0 0 0 0 0 0 0 0 0 863 1680
proc3 22 815 1230508 0 12217802 486762 304 405845161 1037854 2 2 0 0 4
2 2 0 0 559224 48069 14 0 66205

If the th line has large numbers, add more nfsd's - Rajiv Bendale.

Looking for wireless networks:

# iwlist scan

If you want a continuous monitor (more sophisticated) look into 
"kismet" http://www.kismetwireless.net/.

iwconfig

Slow down rate of linux software raid resync after disk failure and
replacement:
> reduce the numbers in:
> /proc/sys/dev/raid/speed_limit_max
> /proc/sys/dev/raid/speed_limit_min

From a guy on oclug:
For best bet on compatibility with Linux go with a Prism based card.
802.11b will be PrismII or PrismIII, and 802.11g will be PrismGT.

cat /proc/fs/nfs/exports

On getting good performance out of synchronous NFS:

The best way to get good nfs performance with a sync export is to use
a data-journalling filesystem, and have the journal somewhere fast.

I use ext3.  I believe that some other filesystems support data
journalling, but I don't know the details.

An ext3 filesystem exported with "data=journal" is significantly
faster that with the default "data=ordered".

With ext3, it also helps to export with the "no_wdelay" export
option.  I'm not entirely sure why this helps, and I haven't confirmed
that it is still the case (I discovered this at least 2 years ago) but
it is worth checking in your environment.

Ext3 supports having the journal on a separate device.  If you put it
on a fast device, you will get faster response.

One option is to buy an NV-RAM card and put the journal on that.
www.umem.com is one supplier that I know of.  There are doubtlessly
others.

Another option is to put it on a separate drive, or preferably a
separate mirrored pair.  As journal writes are sequential, you won't
suffer seek time on the journal device and it will be quite fast.

My typical fileserver configuration is to have a pair of small (36gig
is the smallest I can easily get) 15K SCSI drives as a mirrored pair,
and to put root, swap, and a journal partition on this.
Then a separate box contains a dozen or so drives of whatever size I
want in a raid5 configuration.  The ext3 filesystem is on the raid5
array, the journal is on the mirrored pair with root and swap.  There
is almost no IO to root and swap, so the journal gets the drives
almost to itself.
This configuration is adequately fast.

NeilBrown

> You will need to stop Evolution backend tasks and the GConf daemon
> before copying your data. On *both* machines (syncing the data to disk
> on the source machine and not using old data on target machine):
> 
> $ evolution --force-shutdown
> $ gconftools-2 --shutdown

Doh! Make that 'gconftool-2'. Sorry for any confusion this might have
caused...

...guenther

iptstate is pretty cool.  It should current iptables-related connections,
apparently.  It comes with FC3 and Devil Linux 1.2, but not FC2 or RHEL 3.

On migrating linux NFS servers:

Yes - existing clients will still use the 'default' filehandle type. Any 
new clients will use the fsid= option type. You have to make sure all 
clients have remounted the file systems before moving the disks to a new 
host.

James Pearson

David Dougall wrote:
> So, you are saying that if I add an fsid= option into the exports, the
> server will work with both that filehandle and the default major/minor
> number one?
> --David Dougall
> 
> 
> On Mon, 13 Dec 2004, James Pearson wrote:
> 
> 
>>David Dougall wrote:
>>
>>>I am moving filesystems/disks to new NFS servers.  They are
going to
>>>change major and minor numbers.  Is there any way I can
prevent stale
>>>mounts with the fsid= option in the exports?
>>>I have tried this on a test environment, but when I try to
create an fsid=
>>>option, it completely changes the filehandle format in the
nfs packet.  Am
>>>I missing something, or are the default and forced fsid options
>>>incompatible?
>>
>>I believe using the fsid= option does use a different file handle id
>>type, so you can't directly do what you want with the fsid option.
>>
>>However, you can re-export a file system with an added fsid= option -
>>any existing mounts will still use the 'previous' file handle
type based
>>on the device id, but new mounts will use the new fsid file handle. You
>>then will have to make sure all existing clients remount the
file system
>>before moving the disks.
>>
>>I've done this before - but I had to make the export change well in
>>advance, so that I could make sure all existing clients remounted the
>>file system before the move.
>>
>>James Pearson

I would start by turning on the RPC debugging:
Try
    sysctl -w sunrpc.rpc_debug=3

(or "echo 3 >/proc/sys/sunrpc/rpc_debug" - which does the same thing).

The first thing to do is to look for "RPC: %4d received reply" and try
to match the pid to the request pid. Also look for timeouts.

> what are the other possible debug settings besides '3'
> ??

See the RPCDBG_* values in include/linux/sunrpc/debug.h: they define
the bitmask of "dprintk()"s that can be turned on.

udevinfo -a -p (dir)
helpful in constructing /dev entries

From Duncan:

We've updated the certificate for RHN so you'll need to
run the following if your system isn't checking in:

  rpm -e rhns-ca-cert-1.0-1
  rpm -i http://rhn.nacs.uci.edu/pub/rhns-ca-cert-1.0-3.noarch.rpm

Backing up gconf2:

gconftool-2 --dump 

Shutting down gconf2:

gconftool-2 --shutdown

Firstly, you should always use the '-c' flag when measuring NFS client
performance. Otherwise, you are basically just measuring the speed with
which your machine can write to the local page cache (which may indeed
explain your "knee" at 1MB here - that would be where the client starts
to force a flush to disk).

Installing grub on a hard disk:
# grub
Grub>device (hd0) /dev/sdb (/dev/hdb for ide)
Grub>root (hd0,0) and then:
Grub>setup (hd0)

Also:
grub-install /dev/sda

Blocking a host with iptables:
iptables -A INPUT -s yahoo.com -j DROP

On manipulating source rpm's:

See http://postfix.wl0.org/ftp/rpmdiff/
which includes

rpmdiff         A script which allows you to do a diff on 2 similiar
                source rpms.
rpmextract      A wrapper around rpm2cpio for extracting the contents
                of a src rpm.
rpmpatch        For patching an old src rpm to provide a new src rpm.

I moved my home directory from /Dcs/staff/strombrg to /Dcs/seki/strombrg.
Evolution lost track of where its addressbook was.  I was able to fix
it by going into gconf-editor, and editing apps/evolution/addressbook.
There was a value that had memorized a hard path to addressbook in my
old home directory.

Also had to vi my .gtkrc to change the hard path...

Perhaps why we get stale handles on reboots:

On 2.4 kernels it is not 'ornamental' - if the server reboots without an 
entry for a client mount in /var/lib/nfs/rmtab, then the client will get 
a stale NFS file handle when the server reboots. This can happen if 
umount fails e.g. the mount point is busy - rpc.mountd on the server 
removes the entry from /var/lib/nfs/rmtab, but umount fails, so the file 
system remains mounted.

"A CPU chip containing multiple hyper-threaded processing elements is 
counted as a single CPU"

http://www.redhat.com/software/rhel/comparison/

> Normally, on a RHEL system, you just go into /user/share/ssl/certs/ and
>
> type:
> 
> make whatever.pem

go to /usr/local/share/doc/dovecot (on FreeBSD), edit example 
dovecot-openssl.cnf
for your needs and run mkcert.sh

From Duncan, 2005-03-03:

  In order to eliminate the redundancy inherent in providing a separate 
package for the kernel source code when that source code already exists 
in the kernel's .src.rpm file, Red Hat Enterprise Linux 4 no longer 
includes the kernel-source package. Users that require access to the 
kernel sources can find them in the kernel .src.rpm file. To create an 
exploded source tree from this file, perform the following steps (note 
that <version> refers to the version specification for your 
currently-running kernel):

 1.  Obtain the kernel-<version>.src.rpm file from one of the following 
      sources:
    o The SRPMS directory on the appropriate "SRPMS" CD iso image
    o The FTP site where you got the kernel package
    o By running the following command:
      up2date --get-source kernel

 2.  Install kernel-<version>.src.rpm (given the default RPM
 configuration,
     the files this package contains will be written to /usr/src/redhat/)

 3.  Change directory to /usr/src/redhat/SPECS/, and issue the following 
     command:
       rpmbuild -bp --target=<arch> kernel.spec
       (Where <arch> is the desired target architecture.)

      On a default RPM configuration, the kernel tree will be located in 
      /usr/src/redhat/BUILD/.

 4.  In resulting tree, the configurations for the specific kernels shipped 
     in Red Hat Enterprise Linux 4 are in the /configs/ directory. For 
     example, the i686 SMP configuration file is named 
     /configs/kernel-<version>-i686-smp.config. Issue the following
     command
     to place the desired configuration file in the proper place for building:
         cp <desired-file> ./.config

 5.  Issue the following command:
         make oldconfig

  You can then proceed as usual.


Note

  An exploded source tree is not required to build kernel modules against 
  the currently in-use kernel.

  For example, to build the foo.ko module, create the following file (named 
  Makefile) in the directory containing the foo.c file:


   obj-m    := foo.o

   KDIR    := /lib/modules/$(shell uname -r)/build
   PWD    := $(shell pwd)

   default:
       $(MAKE) -C $(KDIR) SUBDIRS=$(PWD) modules

  

     Issue the make command to build the foo.ko module.

FWIW Red Hat appears to officially support 8TB ext3 filesystems on
Red Hat Enterprise Linux 4:

"Ext3 scalability: Dynamic file system expansion and file system sizes
up to 8TB are now supported."

http://www.redhat.com/software/rhel/features/

From Duncan:

RHEL 4 installs with a default setting that rejects 
remote connections to the local X server despite the 
xhost or mxconns settings. 

To turn this behavior off, edit /etc/X11/gdm/gdm.conf
and make sure this DisallowTCP is set to false.

Nice talk that covers the relationships between different RAID/Volume
management software
http://people.redhat.com/agk/talks/FOSDEM_2005/

On speeding up huge ext3 directories:

[Look for feature "dir_index" on the "Filesystem features:" line in
"/sbin/tune2fs -l <device>" output.]