Hi,

I just migrated from OpenSlug 2.7 to the current Debian version.

My first performance test with samba shares showed that my current
installation is significantly slower that before.

As far as I can see, I configured samba as I did before when using
OpenSlug.

Before I hat a read performance of about 4,5 kb per second.
Now it's only 2,8 to 3,0 kb per second.

Any idea what's wrong? Any hints where to find some hints?

I don't know, but I believe I used ext2 as file system before and now
it is ext3. Does this make the trick?

Ah by the way: in boths setups I used a mdadm driven raid 1 system
(same hardware)

Any ideas are highly appreciated.

Thanks in advance
Markus

Oh,

of course it's MB per sec ;-)

Markus

markusingendahl schrieb:
>
> Hi,
>
> I just migrated from OpenSlug 2.7 to the current Debian version.
>
> My first performance test with samba shares showed that my current
> installation is significantly slower that before.
>
> As far as I can see, I configured samba as I did before when using
> OpenSlug.
>
> Before I hat a read performance of about 4,5 kb per second.
> Now it's only 2,8 to 3,0 kb per second.
>
> Any idea what's wrong? Any hints where to find some hints?
>
> I don't know, but I believe I used ext2 as file system before and now
> it is ext3. Does this make the trick?
>
> Ah by the way: in boths setups I used a mdadm driven raid 1 system
> (same hardware)
>
> Any ideas are highly appreciated.
>
> Thanks in advance
> Markus
>
>

Hi all,
I recently installed lighttpd un my slug (unslung 6.8) and created
two websites. I use virtual hosting for two sites with two IP-names.
Wonderfull, so far. The server was however still running as root,
which I changed to user www.
I created this user from unslung with the adduser command. I also
use sftp to update the sites from another machine inside my LAN. I
have to be able to update it from outside my LAN, so sftp is a must.
After reboot the new user www was changed: it didn't have a password
anymore. The passwd file just had an x. I searched the wiki and
found that this is normal behaviour, that I should create a user
from the Linksys interface. So I deleted my own www user (and www
group) and created one from the webinterface. I chown'd all the
files in my sites to www.www. Wonderfull! Both my sites are up and
running again.
BUT: the Linksys interface sets the shell for www to /dev/null.
I did not find a moduser command for unslung, so I cannot change the
shell in passwd. So, without a valid shell, I cannot sftp to the
slug with user www. I have to use root, and all the files I put to
my server are owned by root, so the webserver has no access to it. I
first have to chown them before I can start testing.
Any solutions? I'm not in any way an experienced Linux user.
Thanks in advance.
Jos

In article <fnv5ig+bneg@eGroups.com>,
markusingendahl <mark.i@...> wrote:

> Hi,
>
> I just migrated from OpenSlug 2.7 to the current Debian version.
>
> My first performance test with samba shares showed that my current
> installation is significantly slower that before.
>
> As far as I can see, I configured samba as I did before when using
> OpenSlug.
>
> Before I hat a read performance of about 4,5 kb per second.
> Now it's only 2,8 to 3,0 kb per second.
>
> Any idea what's wrong? Any hints where to find some hints?
>
> I don't know, but I believe I used ext2 as file system before and now
> it is ext3. Does this make the trick?
>
> Ah by the way: in boths setups I used a mdadm driven raid 1 system
> (same hardware)

Sorry, no ideas, but firstly I assume you mean MB (Megabytes) per
second?

I did not know what speed to expect so I transferred a big file (just
under 2GB) from the slug to a WinXP box using Samba.

terminus:/home/bob# uname -a
Linux terminus 2.6.18-4-ixp4xx #1 Tue Mar 27 18:01:56 BST 2007 armv5tel
GNU/Linux

Samba is version 3.0.24-6.

5 minutes 33 seconds to transfer 1,967,128,576 bytes, which is 5,907,293
bytes/second or 5.9 Megabytes (MB)/sec (or approximately 47 Megabit
(Mb)/sec)

So, yes, 2.8 to 3MB per second is perhaps a bit slow.

The Windoze box is an old one running XP SP1.

--
Bob Cox.  Stoke Gifford, near Bristol, UK.
Registered user #445000 with the Linux Counter: http://counter.li.org/
Debian on NSLU2 Slug: http://bobcox.com/slug/

vi /etc/passswd and change the shell?

Ben

On Feb 1, 2008 2:08 PM, joswennmacker <joswennmacker@...> wrote:
>
>
>
>
>
>
> Hi all,
>  I recently installed lighttpd un my slug (unslung 6.8) and created
>  two websites. I use virtual hosting for two sites with two IP-names.
>  Wonderfull, so far. The server was however still running as root,
>  which I changed to user www.
>  I created this user from unslung with the adduser command. I also
>  use sftp to update the sites from another machine inside my LAN. I
>  have to be able to update it from outside my LAN, so sftp is a must.
>  After reboot the new user www was changed: it didn't have a password
>  anymore. The passwd file just had an x. I searched the wiki and
>  found that this is normal behaviour, that I should create a user
>  from the Linksys interface. So I deleted my own www user (and www
>  group) and created one from the webinterface. I chown'd all the
>  files in my sites to www.www. Wonderfull! Both my sites are up and
>  running again.
>  BUT: the Linksys interface sets the shell for www to /dev/null.
>  I did not find a moduser command for unslung, so I cannot change the
>  shell in passwd. So, without a valid shell, I cannot sftp to the
>  slug with user www. I have to use root, and all the files I put to
>  my server are owned by root, so the webserver has no access to it. I
>  first have to chown them before I can start testing.
>  Any solutions? I'm not in any way an experienced Linux user.
>  Thanks in advance.
>  Jos
>
>



--
"A Scientist will earn a living by taking a really difficult problem
and spends many years solving it, an engineer earns a living by
finding really difficult problems and side stepping them"

Brilliant. Thanks. Can this be used alongside the Orange DCHP Server
or will I need to disable this in the Orange Livebox?

--- In nslu2-linux@yahoogroups.com, Frank Reijn <fnijer@...> wrote:
>
> Hi Raspden,
>
> No it can not..... The orange box I mean  but.........
>
> Why not creating your own DNS server  :-)
> Although it is a small box, it has a lot of power !!!!!!
>
> http://www.nslu2-linux.org/wiki/Optware/DnsMasq
>
> Be carefull now...  wrongly setting up internal DNS can mess up your
outgoing internet connection.
>
> Cheers,
>
>   Frank
>
>
>
> raspden <raspden@...> wrote:                               Thanks.
That's cleared things up nicely. I didn't realise I still
>  needed to use the hosts file.
>
>  The local network is just my basic home network. It's run from an
>  Orange Livebox which has a DHCP server. All very simple. Have had a
>  look in the livebox setup and can find no sign of DNS Servers. There
>  is dynamic DNS, but I don't think that can be used for this
purpose(???).
>
>  Does anyone know if the DNS Servers can be resolved on the Livebox?
>
>  If not I'll just use hosts.
>
>  Thanks again.
>
>  --- In nslu2-linux@yahoogroups.com, "Scott Ruckh" <sruckh@> wrote:
>  >
>  >
>  > This is what you said Brian Wood
>  > > Frank Reijn wrote:
>  > >>
>  > >>
>  > >> You do have two questions here.
>  > >>
>  > >> 1 - what and how is DNS
>  > >>
>  > >> 2 - Virtual hosting
>  > >>
>  > >>
>  > >> I give you the answer on Q 1.
>  > >>
>  > >> If you type http://192.168.1.77:8081   your browser knows which
>  machine
>  > >> to query because it uses the ip adress to reach your slug.
>  > >>
>  > >> if you type http://pma.rhazzzz.com:8081/
>  > >> <http://pma.rhazzzz.com:8081/>   your browser will query the
>  configured
>  > >> DNS server and ask it for the relevant ipaddress.
>  > >> As you stated this is a FAKE  url and  the DNS server does not
>  have an
>  > >> entry for this and your journey on the network stops....
>  > >>
>  > >> Eighter you have a local network DNS and put the pma.rhazzzz.com
>  in , or
>  > >> you do need a host file on EVERY computer to translate
>  pma.rhazzzz.com
>  > >> into
>  > >> 192.168.1.77.
>  > >>
>  > >> You just did your first 2 baby steps on the network  :-)
>  > >>
>  > >> For the second question I can not help you... sorry...
>  > >
>  > > I can, perhaps:
>  > >
>  > > Normally and originally a web server had a single address and
hosted a
>  > > single site.
>  > >
>  > > Virtual hosting allows you to host more than one site using a
single
>  > > machine.
>  > >
>  > > One way to do this is to run multiple web servers on different
ports,
>  > > but this doesn't work well because most people and systems expect
>  a web
>  > > server to be on a certain port.
>  > >
>  > > "Name virtual hosting" allows multiple host names to all resolve
>  to the
>  > > same IP address. The server knows by what name it was called,
and will
>  > > serve up the appropriate site based on this.
>  > >
>  > > The problem, of course, is when somebody calls the site via its
>  > > numerical address, not a name. Now the server has no way to
know what
>  > > site is wanted and will serve up only the "default" site.
>  > >
>  > > But as long as the sites are callled by name, and each name has
a DNS
>  > > entry that points to the same server, the server can decide
which site
>  > > to serve up based on what name it was called by.
>  > >
>  > > Name virtual hosting has to be enabled in the server, and a
separate
>  > > configuration has to be made for each named site ("named virtual
>  host").
>  > >
>  > > Hope that helps.
>  > >
>  > For the original poster, I will say this again (especially knowing
>  this is
>  > for your internal network only).
>  >
>  > If your client (the one running the web browser) is a windows box:
>  > Open up (in your favorite editor)
>  %SystemRoot%\system32\drivers\etc\hosts,
>  > most commonly, C:\WINDOWS\system32\drivers\etc\hosts, and add an
>  entry for
>  > your virtual host your are creating for your http server.  Possibly
>  > something like this:
>  >
>  > 192.168.1.77 pma pma.rhazzzz.com
>  >
>  > On a *NIX type client, the hosts file is typically located here:
>  /etc/hosts.
>  >
>  > Again, add an entry to the hosts file, using the exact same syntax
>  as above.
>  >
>  > This edit of the hosts file will be necessary for every client.
>  >
>  > If you are in control of a DNS server on your network, then obviously
>  > adding possibly a CNAME record for that name would be your best
option.
>  > If you don't maintain a DNS server for your local network, then
use the
>  > hosts file as mentioned above.
>  >
>  > Now from that machine, you should be able to do things like 'ping
>  > pma.rhazzzz.com'.  This assumes that your name look-up is using the
>  hosts
>  > file (which is by default true on most Operating Systems).
>  >
>  > Now that the name resolution problem is resolved, it is now a
matter of
>  > getting your virtual host configuration set up in your lighttpd.conf
>  file
>  > (I have already posted and example of this, so you should be ready
>  to go).
>  >
>
>
>
>
>
>
> ---------------------------------
> Never miss a thing.   Make Yahoo your homepage.
>

Hi

First post (sorry a bit of a long one) here so please be gentle.

Up front, I would like to thank everybody who contribute to this group
and its related sites, and hope that it continues. Without stumbling
into this world I would not have purchased a slug or a buffalo router
(currently running tomato f/w), knowing that they could be made to do
other things.

I have no practical linux experience, but have worked on unix (IBM
RS6000 AIX), so know roughly how things hang together, and am not
afraid of the command line.

I purchased a slug/usb/500gb hdd a few months ago, and have just
gotten around to actually playing with it.

I have been reading through the wiki and this group, and after
reviewing the firmware comparison table, I opted to try and install
the latest OpenWRT firmware. It looks to be lean and mean, and with
the encouragement from Rod's Christmas 2007 posting, seemed to be the
way to go.

My usage initially for the slug, is to be a basic file server running
SAMBA for the windows pcs/laptop, and NFS for the unix based
satellite/cable/media boxes I have. I hope to install a full copy of
linux on my dell laptop, but just waiting to see if the ati card issue
settles down.

Now for my problem:-

I installed OpenWrt Kamikaze 7.09 on my slug and got it connected to
my network without too much pain.
The x-wrt gui is a nice comfort to use after the initial install, and
helps to give some confidence to first timers like myself.
Made a few changes, and can connect with Putty.
I decided to target NFS as my first application to install


First attempt

Ran ipkg update
installed the filesystem management tools
installed the nfs server pkgs from the default feeds
ran the nfs startup script, and thought everything was working
tried to mount something from my media player, but no luck
looked back in dmesg and saw that nfs processes shutdown after 90secs
eventually realised that portmap was not running
when I started it manually it fell over on a seg fault
removed/reinstalled but no change
went back to read some more


Second attempt

removed all pkgs installed during first attempt
created /opt dir
edited /etc/ipkg.conf and installed optware pkg installer
added /opt/bin, /opt/sbin to default path in /etc/profile
ran ipkg update
installed portmap pkg and checked that it started okay
from reading I understood that the nfs-utils pkg is what I want for my
nfs server
pkg did not show up on ipkg list, so after searching on wike found
that version 1.1.1 existed in nslu2 feed
added feed
http://ipkg.nslu2-linux.org/feeds/optware/nslu2/cross/unstable/ to
ipkg.conf file
reran ipkg update, and then installed nfs-utils
at this point everything looked good, so ran poweroff to make sure
everything up to date
tried to start nfs by running script /opt/etc/init.d/S56nfs-utils, but
got 4 errors saying that the 4 binaries where not found
ie: stats, nfsd, exportfs, mountd
when I run them manually using /opt/sbin/xxxx or cd to /opt/sbin and
run them using ./xxxx, I get the same error
each of the files have perms of 755 and the directories are all set at
755 as well
the portmap program is also installed, but it has no problem loading
started reading again, and installed the kernel support pkgs kmod-nfs,
kmod-nfsd, kmod-loop - no change, but are these needed?
copied nfsd to /tmp, renamed it, changed perms, tried to runt it, but
it still gave 'not found' error
ran strings against pgms and noticed references to /lib/ld-linux-so.2
which looked odd
could not find this lib on my system
checked the portmap pgm and it referenced uClibc.so.0
back to reading again, noticed references to ldd output, so I ran that
again the pgms, and it displays correctly for portmap, but not for the
other 4

So my questions are
From my steps detailed above, have I done something wrong or missed
something obvious?
Is the version of nfs-utils the correct one to install for a nfs server?
Lastly, by any chance, is there a problem with the nfs-utils package?

Any direct help, or point me at something else to read, would be
appreciated - not used to this pkg method of installation, so am
unsure where the fault is.

Again, to all who read this, thanks for your time.

Regards
WW

Brilliant. Thanks. Can this be used alongside the Orange DCHP Server
or will I need to disable this in the Orange Livebox?

--- In nslu2-linux@yahoogroups.com, Frank Reijn <fnijer@...> wrote:
>
> Hi Raspden,
>
> No it can not..... The orange box I mean but.........
>
> Why not creating your own DNS server :-)
> Although it is a small box, it has a lot of power !!!!!!
>
> http://www.nslu2-linux.org/wiki/Optware/DnsMasq
>
> Be carefull now... wrongly setting up internal DNS can mess up your
outgoing internet connection.
>
> Cheers,
>
> Frank
>
>
>
> raspden <raspden@...> wrote: Thanks.
That's cleared things up nicely. I didn't realise I still
> needed to use the hosts file.
>
> The local network is just my basic home network. It's run from an
> Orange Livebox which has a DHCP server. All very simple. Have had a
> look in the livebox setup and can find no sign of DNS Servers. There
> is dynamic DNS, but I don't think that can be used for this
purpose(???).
>
> Does anyone know if the DNS Servers can be resolved on the Livebox?
>
> If not I'll just use hosts.
>
> Thanks again.
>
> --- In nslu2-linux@yahoogroups.com, "Scott Ruckh" <sruckh@> wrote:
> >
> >
> > This is what you said Brian Wood
> > > Frank Reijn wrote:
> > >>
> > >>
> > >> You do have two questions here.
> > >>
> > >> 1 - what and how is DNS
> > >>
> > >> 2 - Virtual hosting
> > >>
> > >>
> > >> I give you the answer on Q 1.
> > >>
> > >> If you type http://192.168.1.77:8081 your browser knows which
> machine
> > >> to query because it uses the ip adress to reach your slug.
> > >>
> > >> if you type http://pma.rhazzzz.com:8081/
> > >> <http://pma.rhazzzz.com:8081/> your browser will query the
> configured
> > >> DNS server and ask it for the relevant ipaddress.
> > >> As you stated this is a FAKE url and the DNS server does not
> have an
> > >> entry for this and your journey on the network stops....
> > >>
> > >> Eighter you have a local network DNS and put the pma.rhazzzz.com
> in , or
> > >> you do need a host file on EVERY computer to translate
> pma.rhazzzz.com
> > >> into
> > >> 192.168.1.77.
> > >>
> > >> You just did your first 2 baby steps on the network :-)
> > >>
> > >> For the second question I can not help you... sorry...
> > >
> > > I can, perhaps:
> > >
> > > Normally and originally a web server had a single address and
hosted a
> > > single site.
> > >
> > > Virtual hosting allows you to host more than one site using a
single
> > > machine.
> > >
> > > One way to do this is to run multiple web servers on different
ports,
> > > but this doesn't work well because most people and systems expect
> a web
> > > server to be on a certain port.
> > >
> > > "Name virtual hosting" allows multiple host names to all resolve
> to the
> > > same IP address. The server knows by what name it was called,
and will
> > > serve up the appropriate site based on this.
> > >
> > > The problem, of course, is when somebody calls the site via its
> > > numerical address, not a name. Now the server has no way to
know what
> > > site is wanted and will serve up only the "default" site.
> > >
> > > But as long as the sites are callled by name, and each name has
a DNS
> > > entry that points to the same server, the server can decide
which site
> > > to serve up based on what name it was called by.
> > >
> > > Name virtual hosting has to be enabled in the server, and a
separate
> > > configuration has to be made for each named site ("named virtual
> host").
> > >
> > > Hope that helps.
> > >
> > For the original poster, I will say this again (especially knowing
> this is
> > for your internal network only).
> >
> > If your client (the one running the web browser) is a windows box:
> > Open up (in your favorite editor)
> %SystemRoot%\system32\drivers\etc\hosts,
> > most commonly, C:\WINDOWS\system32\drivers\etc\hosts, and add an
> entry for
> > your virtual host your are creating for your http server. Possibly
> > something like this:
> >
> > 192.168.1.77 pma pma.rhazzzz.com
> >
> > On a *NIX type client, the hosts file is typically located here:
> /etc/hosts.
> >
> > Again, add an entry to the hosts file, using the exact same syntax
> as above.
> >
> > This edit of the hosts file will be necessary for every client.
> >
> > If you are in control of a DNS server on your network, then obviously
> > adding possibly a CNAME record for that name would be your best
option.
> > If you don't maintain a DNS server for your local network, then
use the
> > hosts file as mentioned above.
> >
> > Now from that machine, you should be able to do things like 'ping
> > pma.rhazzzz.com'. This assumes that your name look-up is using the
> hosts
> > file (which is by default true on most Operating Systems).
> >
> > Now that the name resolution problem is resolved, it is now a
matter of
> > getting your virtual host configuration set up in your lighttpd.conf
> file
> > (I have already posted and example of this, so you should be ready
> to go).
> >
>
>
>
>
>
>
> ---------------------------------
> Never miss a thing. Make Yahoo your homepage.
>

    __________|__________                                                      
       \    \_^_/    /                                                         
      __\___{_+_}___/__                                                 
  
           :/   \:  

Regards,

  Frank

  Global Operations Director 

This is what you said joswennmacker
> Hi all,
> I recently installed lighttpd un my slug (unslung 6.8) and created
> two websites. I use virtual hosting for two sites with two IP-names.
> Wonderfull, so far. The server was however still running as root,
> which I changed to user www.
> I created this user from unslung with the adduser command. I also
> use sftp to update the sites from another machine inside my LAN. I
> have to be able to update it from outside my LAN, so sftp is a must.
> After reboot the new user www was changed: it didn't have a password
> anymore. The passwd file just had an x. I searched the wiki and
> found that this is normal behaviour, that I should create a user
> from the Linksys interface. So I deleted my own www user (and www
> group) and created one from the webinterface. I chown'd all the
> files in my sites to www.www. Wonderfull! Both my sites are up and
> running again.
> BUT: the Linksys interface sets the shell for www to /dev/null.
> I did not find a moduser command for unslung, so I cannot change the
> shell in passwd. So, without a valid shell, I cannot sftp to the
> slug with user www. I have to use root, and all the files I put to
> my server are owned by root, so the webserver has no access to it. I
> first have to chown them before I can start testing.
> Any solutions? I'm not in any way an experienced Linux user.
> Thanks in advance.
> Jos
>
In my opinion you probably don't want to give write access to the account
that runs the web server.  Maybe give write access to a few required files
here and there, but otherwise read-only access.  Again, that is just my
own personal practice.  Almost all my files are set to root:lighttpd with
the mode on the files set to 644 and directories set to 755.

If you still want to give write access to your 'www' account, you might
consider installing RSSH (www.pizzashack.org/rssh/) for the shell.  In
this way you can still prevent the 'www' acount from having an interactive
shell, but you can still SCP/SFTP files to your server.  Plus it has all
sorts of configuration settings for configuring the umask and the mode on
files.

I personally don't find it all that inconvenient to upload files using a
normal user account, and using 'sudo' to move my files to the configured
web folders and again using 'sudo' to set the permissions.  It is not like
you are going to be hosting hundreds of sites on an NSLU2.

You might also google the chmod command and consider using the sticky bit
on directories.  I have done this in a few cases on folders outside of the
web directories (like my Music directories), so that I can copy files
using SMB as any configured SAMBA user and still have the group
permissions set correctly so that the web server account can read the
files.

If you are just looking for convenience, just do as another user suggested
and just set your desired shell by editing the /etc/passwd file.  Just
remember in almost every instance secure does not mean convenient.

Ok. Will try that. Will probably mess it all up but worth giving it a go.

--- In nslu2-linux@yahoogroups.com, Frank Reijn <fnijer@...> wrote:
>
> Best in my opinion is to stop the Orange one and setup your own DHCP
and DNS.
> But then do not forget to link to the external internet for the dns
part as well!
>
>   Frank
>
> raspden <raspden@...> wrote:
Brilliant. Thanks. Can this be used alongside the Orange DCHP Server
>  or will I need to disable this in the Orange Livebox?
>
>  --- In nslu2-linux@yahoogroups.com, Frank Reijn <fnijer@> wrote:
>  >
>  > Hi Raspden,
>  >
>  > No it can not..... The orange box I mean  but.........
>  >
>  > Why not creating your own DNS server  :-)
>  > Although it is a small box, it has a lot of power !!!!!!
>  >
>  > http://www.nslu2-linux.org/wiki/Optware/DnsMasq
>  >
>  > Be carefull now...  wrongly setting up internal DNS can mess up your
>  outgoing internet connection.
>  >
>  > Cheers,
>  >
>  >   Frank
>  >
>  >
>  >
>  > raspden <raspden@> wrote:                               Thanks.
>  That's cleared things up nicely. I didn't realise I still
>  >  needed to use the hosts file.
>  >
>  >  The local network is just my basic home network. It's run from an
>  >  Orange Livebox which has a DHCP server. All very simple. Have had a
>  >  look in the livebox setup and can find no sign of DNS Servers. There
>  >  is dynamic DNS, but I don't think that can be used for this
>  purpose(???).
>  >
>  >  Does anyone know if the DNS Servers can be resolved on the Livebox?
>  >
>  >  If not I'll just use hosts.
>  >
>  >  Thanks again.
>  >
>  >  --- In nslu2-linux@yahoogroups.com, "Scott Ruckh" <sruckh@> wrote:
>  >  >
>  >  >
>  >  > This is what you said Brian Wood
>  >  > > Frank Reijn wrote:
>  >  > >>
>  >  > >>
>  >  > >> You do have two questions here.
>  >  > >>
>  >  > >> 1 - what and how is DNS
>  >  > >>
>  >  > >> 2 - Virtual hosting
>  >  > >>
>  >  > >>
>  >  > >> I give you the answer on Q 1.
>  >  > >>
>  >  > >> If you type http://192.168.1.77:8081   your browser knows which
>  >  machine
>  >  > >> to query because it uses the ip adress to reach your slug.
>  >  > >>
>  >  > >> if you type http://pma.rhazzzz.com:8081/
>  >  > >> <http://pma.rhazzzz.com:8081/>   your browser will query the
>  >  configured
>  >  > >> DNS server and ask it for the relevant ipaddress.
>  >  > >> As you stated this is a FAKE  url and  the DNS server does not
>  >  have an
>  >  > >> entry for this and your journey on the network stops....
>  >  > >>
>  >  > >> Eighter you have a local network DNS and put the
pma.rhazzzz.com
>  >  in , or
>  >  > >> you do need a host file on EVERY computer to translate
>  >  pma.rhazzzz.com
>  >  > >> into
>  >  > >> 192.168.1.77.
>  >  > >>
>  >  > >> You just did your first 2 baby steps on the network  :-)
>  >  > >>
>  >  > >> For the second question I can not help you... sorry...
>  >  > >
>  >  > > I can, perhaps:
>  >  > >
>  >  > > Normally and originally a web server had a single address and
>  hosted a
>  >  > > single site.
>  >  > >
>  >  > > Virtual hosting allows you to host more than one site using a
>  single
>  >  > > machine.
>  >  > >
>  >  > > One way to do this is to run multiple web servers on different
>  ports,
>  >  > > but this doesn't work well because most people and systems
expect
>  >  a web
>  >  > > server to be on a certain port.
>  >  > >
>  >  > > "Name virtual hosting" allows multiple host names to all resolve
>  >  to the
>  >  > > same IP address. The server knows by what name it was called,
>  and will
>  >  > > serve up the appropriate site based on this.
>  >  > >
>  >  > > The problem, of course, is when somebody calls the site via its
>  >  > > numerical address, not a name. Now the server has no way to
>  know what
>  >  > > site is wanted and will serve up only the "default" site.
>  >  > >
>  >  > > But as long as the sites are callled by name, and each name has
>  a DNS
>  >  > > entry that points to the same server, the server can decide
>  which site
>  >  > > to serve up based on what name it was called by.
>  >  > >
>  >  > > Name virtual hosting has to be enabled in the server, and a
>  separate
>  >  > > configuration has to be made for each named site ("named virtual
>  >  host").
>  >  > >
>  >  > > Hope that helps.
>  >  > >
>  >  > For the original poster, I will say this again (especially knowing
>  >  this is
>  >  > for your internal network only).
>  >  >
>  >  > If your client (the one running the web browser) is a windows box:
>  >  > Open up (in your favorite editor)
>  >  %SystemRoot%\system32\drivers\etc\hosts,
>  >  > most commonly, C:\WINDOWS\system32\drivers\etc\hosts, and add an
>  >  entry for
>  >  > your virtual host your are creating for your http server.
Possibly
>  >  > something like this:
>  >  >
>  >  > 192.168.1.77 pma pma.rhazzzz.com
>  >  >
>  >  > On a *NIX type client, the hosts file is typically located here:
>  >  /etc/hosts.
>  >  >
>  >  > Again, add an entry to the hosts file, using the exact same syntax
>  >  as above.
>  >  >
>  >  > This edit of the hosts file will be necessary for every client.
>  >  >
>  >  > If you are in control of a DNS server on your network, then
obviously
>  >  > adding possibly a CNAME record for that name would be your best
>  option.
>  >  > If you don't maintain a DNS server for your local network, then
>  use the
>  >  > hosts file as mentioned above.
>  >  >
>  >  > Now from that machine, you should be able to do things like 'ping
>  >  > pma.rhazzzz.com'.  This assumes that your name look-up is
using the
>  >  hosts
>  >  > file (which is by default true on most Operating Systems).
>  >  >
>  >  > Now that the name resolution problem is resolved, it is now a
>  matter of
>  >  > getting your virtual host configuration set up in your
lighttpd.conf
>  >  file
>  >  > (I have already posted and example of this, so you should be ready
>  >  to go).
>  >  >
>  >
>  >
>  >
>  >
>  >
>  >
>  > ---------------------------------
>  > Never miss a thing.   Make Yahoo your homepage.
>  >
>
>
>
>
>
>
>
>
>
>     __________|__________

>        \    \_^_/    /

>       __\___{_+_}___/__

>            :/   \:
>
> Regards,
>
>   Frank
>
>   Global Operations Director
>
>
>
>
>
> ---------------------------------
> Be a better friend, newshound, and know-it-all with Yahoo! Mobile.
Try it now.
>

Hello!
This is something that I've researched to the vanishing point for a large
number of people. A virtual host, on the Linksys NSLU2 can obviously be
done, and many people have explained thusly how that would work out. I do
know enough about Apache, and since it enjoys a lot of memory, it may not
work on the thing.

However if you are going to be running a DHCP service, then it can only be
run from one unit only. This can be either your Orange Box gizmo, or even
your router, or a local Linux box who's connected to that router. Multiple
DHCP servers tend to confuse even the best managed networks.

Mailhosts that is a subject of yet another thread and an even bigger
discussion.
--
Gregg C Levine hansolofalcon@...
"The Force will be with you always." Obi-Wan Kenobi
 
-----Original Message-----
From: nslu2-linux@yahoogroups.com [mailto:nslu2-linux@yahoogroups.com] On
Behalf Of Frank Reijn
Sent: Thursday, January 31, 2008 2:55 PM
To: nslu2-linux@yahoogroups.com
Subject: Re: [nslu2-linux] Re: Setting up virtual hosts on Linksys NSLU2

Hi Raspden,

No it can not..... The orange box I mean  but.........

Why not creating your own DNS server  :-)
Although it is a small box, it has a lot of power !!!!!!

http://www.nslu2-linux.org/wiki/Optware/DnsMasq

Be carefull now...  wrongly setting up internal DNS can mess up your
outgoing internet connection.

Cheers,

  Frank



raspden <raspden@...> wrote:
Thanks. That's cleared things up nicely. I didn't realise I still
needed to use the hosts file.

The local network is just my basic home network. It's run from an
Orange Livebox which has a DHCP server. All very simple. Have had a
look in the livebox setup and can find no sign of DNS Servers. There
is dynamic DNS, but I don't think that can be used for this purpose(???).

Does anyone know if the DNS Servers can be resolved on the Livebox?

If not I'll just use hosts.

Thanks again.

--- In nslu2-linux@yahoogroups.com, "Scott Ruckh" <sruckh@...> wrote:
>
>
> This is what you said Brian Wood
> > Frank Reijn wrote:
> >>
> >>
> >> You do have two questions here.
> >>
> >> 1 - what and how is DNS
> >>
> >> 2 - Virtual hosting
> >>
> >>
> >> I give you the answer on Q 1.
> >>
> >> If you type http://192.168.1.77:8081 your browser knows which
machine
> >> to query because it uses the ip adress to reach your slug.
> >>
> >> if you type http://pma.rhazzzz.com:8081/
> >> <http://pma.rhazzzz.com:8081/> your browser will query the
configured
> >> DNS server and ask it for the relevant ipaddress.
> >> As you stated this is a FAKE url and the DNS server does not
have an
> >> entry for this and your journey on the network stops....
> >>
> >> Eighter you have a local network DNS and put the pma.rhazzzz.com
in , or
> >> you do need a host file on EVERY computer to translate
pma.rhazzzz.com
> >> into
> >> 192.168.1.77.
> >>
> >> You just did your first 2 baby steps on the network :-)
> >>
> >> For the second question I can not help you... sorry...
> >
> > I can, perhaps:
> >
> > Normally and originally a web server had a single address and hosted a
> > single site.
> >
> > Virtual hosting allows you to host more than one site using a single
> > machine.
> >
> > One way to do this is to run multiple web servers on different ports,
> > but this doesn't work well because most people and systems expect
a web
> > server to be on a certain port.
> >
> > "Name virtual hosting" allows multiple host names to all resolve
to the
> > same IP address. The server knows by what name it was called, and will
> > serve up the appropriate site based on this.
> >
> > The problem, of course, is when somebody calls the site via its
> > numerical address, not a name. Now the server has no way to know what
> > site is wanted and will serve up only the "default" site.
> >
> > But as long as the sites are callled by name, and each name has a DNS
> > entry that points to the same server, the server can decide which site
> > to serve up based on what name it was called by.
> >
> > Name virtual hosting has to be enabled in the server, and a separate
> > configuration has to be made for each named site ("named virtual
host").
> >
> > Hope that helps.
> >
> For the original poster, I will say this again (especially knowing
this is
> for your internal network only).
>
> If your client (the one running the web browser) is a windows box:
> Open up (in your favorite editor)
%SystemRoot%\system32\drivers\etc\hosts,
> most commonly, C:\WINDOWS\system32\drivers\etc\hosts, and add an
entry for
> your virtual host your are creating for your http server. Possibly
> something like this:
>
> 192.168.1.77 pma pma.rhazzzz.com
>
> On a *NIX type client, the hosts file is typically located here:
/etc/hosts.
>
> Again, add an entry to the hosts file, using the exact same syntax
as above.
>
> This edit of the hosts file will be necessary for every client.
>
> If you are in control of a DNS server on your network, then obviously
> adding possibly a CNAME record for that name would be your best option.
> If you don't maintain a DNS server for your local network, then use the
> hosts file as mentioned above.
>
> Now from that machine, you should be able to do things like 'ping
> pma.rhazzzz.com'. This assumes that your name look-up is using the
hosts
> file (which is by default true on most Operating Systems).
>
> Now that the name resolution problem is resolved, it is now a matter of
> getting your virtual host configuration set up in your lighttpd.conf
file
> (I have already posted and example of this, so you should be ready
to go).
>

Markus,

On Fri, Feb 01, 2008 at 01:07:28PM -0000, markusingendahl wrote:
> Hi,
>
> I just migrated from OpenSlug 2.7 to the current Debian version.
>
> My first performance test with samba shares showed that my current
> installation is significantly slower that before.
>
> [snip]
>
> I don't know, but I believe I used ext2 as file system before and now
> it is ext3. Does this make the trick?

      That is a simple one to check.  Unmount the ext3 disk and then
mount it as ext2 (mount -t ext2 ....).  This is possible because
ext3 was designed to be ext2 compatible.

      While you are at it, make sure you have the noatime mount
option set for all ext2/ext3 disks.

      Also, Debian has many usually unnecessary services, like a
mail server, running by default.  To reduce swapping (thereby
increasing slug speed), you can turn them off.

John

white_westieb wrote:
> pkg did not show up on ipkg list, so after searching on wike found
> that version 1.1.1 existed in nslu2 feed
> added feed
> http://ipkg.nslu2-linux.org/feeds/optware/nslu2/cross/unstable/ to
> ipkg.conf file
> reran ipkg update, and then installed nfs-utils

You can't run optware/nslu2 packages on openwrt.

There is a specific optware feed for openwrt-ixp4xx.

-- Rod

I'm using my SlugOS/BE 6.10 slug as a backup server, running rsnapshot and I see
lots of error messages like this:
rsync: failed to set times on
"/home/snapshot/.sync/slug/disk1/etc/rc1.d/K09sshd": Function not implemented
(38)

The error messages all relate to symbolic links. Some googling suggests that
there is a new  glibc function, lutimes(), that
implements setting the times on a symlink rather than the file it points to, and
that the error message occurs when there's a
discrepancy between the libs available when rsync is built and when it's
deployed, with kernel support required too.

Source:
http://www.mail-archive.com/jfs-discussion@lists.sourceforge.net/msg01001.html

I installed SlugOS a couple of weeks ago, so everything is using the latest
feeds. My rsync tells me it's version 2.6.9, although
that doesn't tell me when it was compiled, or with what environment. Running ldd
on rsync gives me this:
ldd `which rsync`
         libpopt.so.0 => /usr/lib/libpopt.so.0 (0x40022000)
         libgcc_s.so.1 => /lib/libgcc_s.so.1 (0x40030000)
         libc.so.6 => /lib/libc.so.6 (0x40042000)
         /lib/ld-linux.so.2 (0x40000000)

Does this make sense to anybody? Is this likely to be my problem? What is the
appropriate fix?

Thanks,
Paul

--- In nslu2-linux@yahoogroups.com, Rod Whitby <rod@...> wrote:
>
> white_westieb wrote:
> > pkg did not show up on ipkg list, so after searching on wike found
> > that version 1.1.1 existed in nslu2 feed
> > added feed
> > http://ipkg.nslu2-linux.org/feeds/optware/nslu2/cross/unstable/ to
> > ipkg.conf file
> > reran ipkg update, and then installed nfs-utils
>
> You can't run optware/nslu2 packages on openwrt.
>
> There is a specific optware feed for openwrt-ixp4xx.
>
> -- Rod
>

Rod
Thanks for response - this is one area I am unsure of.

Looking at that feed again, there is no nfs server - the comment
against unfs3 pkg is to use nfs-utils, which I did, but it don't work.
Is there an nfs server pkg built to run on Openwrt?
There is the nfs-kernel-server.1.0.7 pkg in openwrt/kamikaze-7.07, but
I think its only an nfs2 server?
Any suggestions?

Regards
WW

white_westieb wrote:
> Looking at that feed again, there is no nfs server - the comment
> against unfs3 pkg is to use nfs-utils, which I did, but it don't work.
> Is there an nfs server pkg built to run on Openwrt?

No idea.  None of my openwrt boxes have attached storage, so I've never
needed an NFS server for them.

-- Rod

This is what you said Brendan Reid
> hi all
>
> env:
> slug (266)
> unslung
> used file server and backup box (rsnapshot/rsync) mostly
> firmware:  Unslung 6.8
>
> noticed a couple of days ago that slug time was one and a bit hours ahead
> of
> time on my pc (cron job starting earlier than expected)
>
> set the date to correct time using "date -s"
>
> checked again tonight and it is 1 hour and 15mins ahead again
>
> thoughts?
>
> bren
>

http://www.nslu2-linux.org/wiki/HowTo/FixTheClockUsingAdjtimexAndNtpclient

This is what you said Gregg C Levine
> Hello!
> This is something that I've researched to the vanishing point for a large
> number of people. A virtual host, on the Linksys NSLU2 can obviously be
> done, and many people have explained thusly how that would work out. I do
> know enough about Apache, and since it enjoys a lot of memory, it may not
> work on the thing.
>
> However if you are going to be running a DHCP service, then it can only be
> run from one unit only. This can be either your Orange Box gizmo, or even
> your router, or a local Linux box who's connected to that router. Multiple
> DHCP servers tend to confuse even the best managed networks.
>
> Mailhosts that is a subject of yet another thread and an even bigger
> discussion.
> --

I have the luxury of having a dedicated Linux box that acts as firewall,
router, DNS, DHCP server, etc.  Although I run CentOS there are several
self contained distributions made just for the purpose of being an
internet gateway.  They have web interfaces for configuration in hopes to
simplify the setup.

This is not entirely practical for everyone, but it is definitely a
solution that has worked well for me (including a mail host gateway, which
the NSLU2 w/sendmail are configured to use).

This is what you said Brendan Reid
> ok i am now struggling with wiki entry here
>
> http://www.nslu2-linux.org/wiki/HowTo/FixTheClockUsingAdjtimexAndNtpclient
>
> which gives a recipe for unslung 6.8 (which is what I am using)
> after following the instructions at the top i have....(actually this page
> seems a real mix of new and outdated info)
>
> ________________________________
> # pwd
> /opt/etc/init.d
> # ls -l
> total 40
> -rwxr-xr-x 1 root root  73 Sep 23 13:35 S05bash
> -rwxr-xr-x 1 root root  83 Jun 21  2007 S05coreutils
> -rwxr-xr-x 1 root root 222 Feb  1 23:57 S10ntp
> lrwxrwxrwx 1 root root  20 Sep 30 03:36 S30iptables ->
> /opt/etc/iptables.sh
> -rwxr-xr-x 1 root root 159 Oct  4 20:32 S33ddclient
> lrwxrwxrwx 1 root root  25 Oct  5 18:15 S35mythttpd ->
> /opt/sbin/webserverreboot
> -rwxr-xr-x 1 root root 101 Oct 10 20:11 S36gnump3d
> -rwxr-xr-x 1 root root 164 Sep  8 08:45 S40sshd
> -rwxr-xr-x 1 root root 108 Jul 17  2006 S55portmap
> -rwxr-xr-x 1 root root 476 Sep 18 21:54 S56nfs-utils
> -rwxr-xr-x 1 root root 273 Sep  9 10:12 S57rsyncd
> -rwxr-xr-x 1 root root 313 Sep 24 23:18 S60mt-daapd
> __________________________________________________________
>
> if i now manually run S10ntp as below (i double checked I am using the
> exact
> script on the wiki page) i get the following screen output and the script
> seems caught in a loop (i have to CTRL-C out of it......)
>
> # ./S10ntp
>     mode:         16384
> -o  offset:       0
> -f  frequency:    0
>     maxerror:     16384000
>     esterror:     16384000
>     status:       64 ( UNSYNC )
> -p  timeconstant: 2
>     precision:    1
>     tolerance:    33554432
> -t  tick:         10000
>     time.tv_sec:  1201911669
>     time.tv_usec: 244468
>     return value: 5 (clock not synchronized)
>
> __________________________________________
>
> if i manually try to run the 6th line of the script i get the same.
> if i ping pool.ntp.org i get no problem
>
> ideas?
>
> bren
>
> ____________
> here is my copy of script
> ______________
> #!/bin/sh
> if [ -n "`pidof ntpclient`" ]; then
>       /bin/killall ntpclient 2>/dev/null
> fi
> /opt/sbin/adjtimex -t 10000
> /opt/bin/ntpclient -s -l -h pool.ntp.org > /var/log/ntp.log
> hwclock --systohc --utc
>

Actually that second recipe never worked for me.

I run the the /unslunc/rc.rstxxxx script at boot, and then I just have a
cron job the syncs the clock every 20 minutes.

I never could get the ntpclient to keep time when running in daemon mode
(as suggested in the recipe).

More or less I only do the top section.

On Monday 28 January 2008 04:40:31 Markus Ingendahl wrote:
> Hi,
>
> I managed to get it up and running after 3 hours!
>
> The reason: I followed the following guide to prepare to run Debian from
> a flash drive:
> http://www.cyrius.com/debian/nslu2/linux-on-flash.html
>
> That's what caused to OOM.
> I just skipped these "optimizations" and no OOM any more.
>
> Any idea why?
>

I had a similar problem when trying to install debian with a 256 MB swap.
I ended up increasing it to 1 GB, with which it finally worked.

There's a post about the linux kernel overcommitting it's swap (check my
thread from a few days ago - there's a link to the issue).  I think this is
what's happening..

One of the optimisations given on the cyrius page is reducing the swap usage,
which I can see being a problem on the slug because there's not much memory
anyway.

On 01/02/2008, Scott Ruckh <sruckh@...> wrote:
>
>
>
>
>
>
> This is what you said Brendan Reid
>
>  > ok i am now struggling with wiki entry here
>  >
>  > http://www.nslu2-linux.org/wiki/HowTo/FixTheClockUsingAdjtimexAndNtpclient
>  >
>  > which gives a recipe for unslung 6.8 (which is what I am using)
>  > after following the instructions at the top i have....(actually this page
>  > seems a real mix of new and outdated info)
>  >
>  > ________________________________
>  > # pwd
>  > /opt/etc/init.d
>  > # ls -l
>  > total 40
>  > -rwxr-xr-x 1 root root  73 Sep 23 13:35 S05bash
>  > -rwxr-xr-x 1 root root  83 Jun 21  2007 S05coreutils
>  > -rwxr-xr-x 1 root root 222 Feb  1 23:57 S10ntp
>  > lrwxrwxrwx 1 root root  20 Sep 30 03:36 S30iptables ->
>  > /opt/etc/iptables.sh
>  > -rwxr-xr-x 1 root root 159 Oct  4 20:32 S33ddclient
>  > lrwxrwxrwx 1 root root  25 Oct  5 18:15 S35mythttpd ->
>  > /opt/sbin/webserverreboot
>  > -rwxr-xr-x 1 root root 101 Oct 10 20:11 S36gnump3d
>  > -rwxr-xr-x 1 root root 164 Sep  8 08:45 S40sshd
>  > -rwxr-xr-x 1 root root 108 Jul 17  2006 S55portmap
>  > -rwxr-xr-x 1 root root 476 Sep 18 21:54 S56nfs-utils
>  > -rwxr-xr-x 1 root root 273 Sep  9 10:12 S57rsyncd
>  > -rwxr-xr-x 1 root root 313 Sep 24 23:18 S60mt-daapd
>  > __________________________________________________________
>  >
>  > if i now manually run S10ntp as below (i double checked I am using the
>  > exact
>  > script on the wiki page) i get the following screen output and the script
>  > seems caught in a loop (i have to CTRL-C out of it......)
>  >
>  > # ./S10ntp
>  >     mode:         16384
>  > -o  offset:       0
>  > -f  frequency:    0
>  >     maxerror:     16384000
>  >     esterror:     16384000
>  >     status:       64 ( UNSYNC )
>  > -p  timeconstant: 2
>  >     precision:    1
>  >     tolerance:    33554432
>  > -t  tick:         10000
>  >     time.tv_sec:  1201911669
>  >     time.tv_usec: 244468
>  >     return value: 5 (clock not synchronized)
>  >
>  > __________________________________________
>  >
>  > if i manually try to run the 6th line of the script i get the same.
>  > if i ping pool.ntp.org i get no problem
>  >
>  > ideas?
>  >
>  > bren
>  >
>  > ____________
>  > here is my copy of script
>  > ______________
>  > #!/bin/sh
>  > if [ -n "`pidof ntpclient`" ]; then
>  >       /bin/killall ntpclient 2>/dev/null
>  > fi
>  > /opt/sbin/adjtimex -t 10000
>  > /opt/bin/ntpclient -s -l -h pool.ntp.org > /var/log/ntp.log
>  > hwclock --systohc --utc
>  >
>
>  Actually that second recipe never worked for me.
>
>  I run the the /unslunc/rc.rstxxxx script at boot, and then I just have a
>  cron job the syncs the clock every 20 minutes.
>
>  I never could get the ntpclient to keep time when running in daemon mode
>  (as suggested in the recipe).
>
>  More or less I only do the top section.


now i am confused!!!!!!

either i use the script at the top plus cron (which uses ntpclient)
or i use the diversion scripts at the bottom (which uses ntpclient)

your post says you only use the top half of what?   An orange?
if you are only running the first script /unslung/rc.rstxxxx the only
command that changes the clock is

  /usr/sbin/hwclock --adjust

but where is this adjustment coming from and relative to what?

as in my original post when i run
                /opt/bin/ntpclient -s -l -h pool.ntp.org
it doesnt work..

am i being thick!
probably

bren



bren

I have worked up a semiconductor parts list for the NSLU2 which I
would like to add as a topic to the Information Pages. The top level
is password-protected, so how do I do this?

Thanks

nigelspon wrote:
> I have worked up a semiconductor parts list for the NSLU2 which I
> would like to add as a topic to the Information Pages. The top level
> is password-protected, so how do I do this?

Create the page, with all it's content in place, and then just mail the
URL of the page to this list, and the location where you would like it
inserted in the parent page.

-- Rod

If you are using an NSLU2 as a dedicated firewall, please contact me
or discuss here.

I want to know about your set-up and how the NSLU2 performs. I have a
router/firewall that directs port 80 traffic to a Slug running
Lighttpd, MySQL, and PHP. The Web server has a static IP and it is
being pummelled by bad guys. I've used up all space for IP rules in my
router/firewall to block the villains and I need to move to a more
powerful solution. I assume the solution will be "iptables" but I have
no experience with it.

This is what you said Brendan Reid
> On 01/02/2008, Scott Ruckh <sruckh@...> wrote:
>>
>>
>>
>>
>>
>>
>> This is what you said Brendan Reid
>>
>>  > ok i am now struggling with wiki entry here
>>  >
>>  >
http://www.nslu2-linux.org/wiki/HowTo/FixTheClockUsingAdjtimexAndNtpclient

The Wiki looks like two different configurations to me.  So when I said
use the one on the top, I meant use the one that looks like option 1.

In /opt/etc/crontab I have the following entry:
0/20 * * * * root /opt/bin/ntpclient -s -l -h a.b.c.d > /var/log/ntp.log
&& hwclock --systohc --utc

;where a.b.c.d is the IP of the NTP server that is on my local network.

for /unslung/rc.rstimezone I have:

#! /bin/ash
/usr/sbin/Set_TimeZone >/dev/null
/usr/sbin/hwclock --adjust
/usr/sbin/hwclock --hctosys >/dev/null
return 0
# EOF - include this line

probably not needed, but I have this for /opt/etc/init.d/S10ntpclient :

#!/bin/sh
if [ -n "`pidof ntpclient`" ]; then
         /bin/killall ntpclient 2>/dev/null
fi
/opt/sbin/adjtimex -t 10000
/opt/bin/ntpclient -s -l -h a.b.c.d > /var/log/ntp.log
/usr/sbin/hwclock --systohc --utc

exit 0

All of that other stuff at the bottom of the Wiki did not work for me.

Either my clock got too skewed for ntpclient to sync, or this just did not
work for me:

/opt/bin/ntpclient -i 600 -l -h a.b.c.d >> /var/log/ntp.log &

So I just opted for the ugly hack, which is to sync to my internal NTP
server every 20 minutes.

1. Install ntpclient: ipkg install ntpclient
2. Add the following line to /etc/crontab: (it's all on one line)

1 * * * * root /opt/bin/ntpclient -s -l -h 0.pool.ntp.org >
/var/log/ntp.log && hwclock --systohc --utc

That's it.
Roland

--- In nslu2-linux@yahoogroups.com, "Brendan Reid" <brenreid@...> wrote:
>
> On 01/02/2008, Scott Ruckh <sruckh@...> wrote:
> >
> >
> >
> >
> >
> >
> > This is what you said Brendan Reid
> >
> >  > ok i am now struggling with wiki entry here
> >  >
> >  >
http://www.nslu2-linux.org/wiki/HowTo/FixTheClockUsingAdjtimexAndNtpclient
> >  >
> >  > which gives a recipe for unslung 6.8 (which is what I am using)
> >  > after following the instructions at the top i have....(actually
this page
> >  > seems a real mix of new and outdated info)
> >  >
> >  > ________________________________
> >  > # pwd
> >  > /opt/etc/init.d
> >  > # ls -l
> >  > total 40
> >  > -rwxr-xr-x 1 root root  73 Sep 23 13:35 S05bash
> >  > -rwxr-xr-x 1 root root  83 Jun 21  2007 S05coreutils
> >  > -rwxr-xr-x 1 root root 222 Feb  1 23:57 S10ntp
> >  > lrwxrwxrwx 1 root root  20 Sep 30 03:36 S30iptables ->
> >  > /opt/etc/iptables.sh
> >  > -rwxr-xr-x 1 root root 159 Oct  4 20:32 S33ddclient
> >  > lrwxrwxrwx 1 root root  25 Oct  5 18:15 S35mythttpd ->
> >  > /opt/sbin/webserverreboot
> >  > -rwxr-xr-x 1 root root 101 Oct 10 20:11 S36gnump3d
> >  > -rwxr-xr-x 1 root root 164 Sep  8 08:45 S40sshd
> >  > -rwxr-xr-x 1 root root 108 Jul 17  2006 S55portmap
> >  > -rwxr-xr-x 1 root root 476 Sep 18 21:54 S56nfs-utils
> >  > -rwxr-xr-x 1 root root 273 Sep  9 10:12 S57rsyncd
> >  > -rwxr-xr-x 1 root root 313 Sep 24 23:18 S60mt-daapd
> >  > __________________________________________________________
> >  >
> >  > if i now manually run S10ntp as below (i double checked I am
using the
> >  > exact
> >  > script on the wiki page) i get the following screen output and
the script
> >  > seems caught in a loop (i have to CTRL-C out of it......)
> >  >
> >  > # ./S10ntp
> >  >     mode:         16384
> >  > -o  offset:       0
> >  > -f  frequency:    0
> >  >     maxerror:     16384000
> >  >     esterror:     16384000
> >  >     status:       64 ( UNSYNC )
> >  > -p  timeconstant: 2
> >  >     precision:    1
> >  >     tolerance:    33554432
> >  > -t  tick:         10000
> >  >     time.tv_sec:  1201911669
> >  >     time.tv_usec: 244468
> >  >     return value: 5 (clock not synchronized)
> >  >
> >  > __________________________________________
> >  >
> >  > if i manually try to run the 6th line of the script i get the same.
> >  > if i ping pool.ntp.org i get no problem
> >  >
> >  > ideas?
> >  >
> >  > bren
> >  >
> >  > ____________
> >  > here is my copy of script
> >  > ______________
> >  > #!/bin/sh
> >  > if [ -n "`pidof ntpclient`" ]; then
> >  >       /bin/killall ntpclient 2>/dev/null
> >  > fi
> >  > /opt/sbin/adjtimex -t 10000
> >  > /opt/bin/ntpclient -s -l -h pool.ntp.org > /var/log/ntp.log
> >  > hwclock --systohc --utc
> >  >
> >
> >  Actually that second recipe never worked for me.
> >
> >  I run the the /unslunc/rc.rstxxxx script at boot, and then I just
have a
> >  cron job the syncs the clock every 20 minutes.
> >
> >  I never could get the ntpclient to keep time when running in
daemon mode
> >  (as suggested in the recipe).
> >
> >  More or less I only do the top section.
>
>
> now i am confused!!!!!!
>
> either i use the script at the top plus cron (which uses ntpclient)
> or i use the diversion scripts at the bottom (which uses ntpclient)
>
> your post says you only use the top half of what?   An orange?
> if you are only running the first script /unslung/rc.rstxxxx the only
> command that changes the clock is
>
>  /usr/sbin/hwclock --adjust
>
> but where is this adjustment coming from and relative to what?
>
> as in my original post when i run
>                /opt/bin/ntpclient -s -l -h pool.ntp.org
> it doesnt work..
>
> am i being thick!
> probably
>
> bren
>
>
>
> bren
>

--- In nslu2-linux@yahoogroups.com, "nigelspon" <nigelspon@...> wrote:
>
> I have worked up a semiconductor parts list for the NSLU2 which I
> would like to add as a topic to the Information Pages. The top level
> is password-protected, so how do I do this?
>
> Thanks
>

Like Rod said, but for the uninitiated, we'll use your info as an
example for others to learn from :-D

If you check out http://www.nslu2-linux.org/wiki/Info/ you'll see a
bunch of Hardware and other stuff.

Type in and go to a URL that doesn't exist, let's say:
http://www.nslu2-linux.org/wiki/Info/NSLU2PartsList

It will ask you if you want to create the topic. Fill out the page
saving along the way. Then post back to this list the URL you created
requesting it be added a locked Area (Info, etc) and section
(Hardware). (NOTE: The HowTo isn't locked, but the Info is.)

That's it. BTW, the Wiki is powered by PmWiki. You can find help on
the NSLU2 Wiki or at the http://www.pmwiki.org site regarding markup
and other elements.

G'night, Mark S.