From pkerrwall at psu.edu  Wed Nov 13 00:45:05 2002
From: pkerrwall at psu.edu (Kerr Wall)
Date: Wed, 13 Nov 2002 00:45:05 -0500
Subject: [Bio-linux] Re: Your clone
In-Reply-To: <5B3EA086-F67F-11D6-B90F-0030657C983C@ceh.ac.uk>
Message-ID: <B9F75111.741B%pkerrwall@psu.edu>

Hi Nick,

On 11/12/02 3:43 PM, "Nicolas Bertrand" <nsb at ceh.ac.uk> wrote:
> I was wondering if you could let me know what video card you purchased
> in the end for your GX260 clone.
> That is important because, video card being the, we may have a GX260
> lying around that could be used to create a perfect image for you to
> rebuild your clone in the short term.
The computer came with a 32MB, ATI, Radeon 7500

> For troubleshooting purposes, could you turn off iptables? Try to ssh
> into fgpclone from 146.186.29.91
> can you connect to fgp clone? Then, ASAP turn iptables back on.
I tried to ssh from 146.186.29.91 and got the same time out problem when I
did this test.

Other than not being able to login except from a 128.118.180.* address, the
computer is running fine and I am able to ssh to one of these other machines
and then go onto the clone.

Should I set up a 'manager' account on one of my other servers to allow one
of you guys to take a look at the files?

Kerr



From pkerrwall at psu.edu  Wed Nov 13 00:49:39 2002
From: pkerrwall at psu.edu (Kerr Wall)
Date: Wed, 13 Nov 2002 00:49:39 -0500
Subject: [Bio-linux] Backing Up
Message-ID: <B9F75223.741D%pkerrwall@psu.edu>

Do any of you have recommendations for a backup setup?

What type of hardware should I purchase and do you have any example scripts
for such a task?

I'll probably limit it to just the 4 linux computers that I have.  I need to
start small and work my way up to getting all of my computers talking to one
another.

Kerr



From mith at ceh.ac.uk  Wed Nov 13 04:40:51 2002
From: mith at ceh.ac.uk (Milo Thurston)
Date: Wed, 13 Nov 2002 09:40:51 +0000
Subject: [Bio-linux] Backing Up 
In-Reply-To: Your message of "Wed, 13 Nov 2002 00:49:39 EST."
             <B9F75223.741D%pkerrwall@psu.edu> 
Message-ID: <200211130940.gAD9ep3K008720@bioinf1.nox.ac.uk>

In message <B9F75223.741D%pkerrwall at psu.edu> you wrote:
>What type of hardware should I purchase and do you have any example scripts
>for such a task?

The cheapest and simplest backup setup is to use dump to store compressed images
of each partition onto a large partition on another disk. I have a couple of machines
containing two disks for this purpose, but it's also possible to dump across the network
using ssh. I can supply a couple of simple scripts to do this, if anyone wants.
This system is cheap and simple for one machine, but does not protect against some sorts
of data loss (theft and fire, for example).
For more elaborate and robust solutions involving tapes, Nicolas will hopefully answer...
Milo.

--
Dr. Milo Thurston, CEH Oxford, Mansfield Road, Oxford, OX1 3SR.
'phone 01865 281658,  fax 01865 281696.
http://www.bioinf.ceh.ac.uk/lab/


From nsb at ceh.ac.uk  Wed Nov 13 07:08:43 2002
From: nsb at ceh.ac.uk (Nicolas Bertrand)
Date: Wed, 13 Nov 2002 12:08:43 +0000
Subject: [Bio-linux] Backing Up
Message-ID: <sdd240d8.072@wpo.nerc.ac.uk>

Kerr,

Here are some of the questions that will need to be addressed

1. What will you backup?
In the case of a bio-linux clone, you probably want to backup soley home The OS can be restored using system imager (minus any system-wide changes you may have done.)

2. When?
A Full weekly backup of home and daily incrementals would be reasonable. 

3. How?
Medium 
DLT Type IV Tapes (4000 and 8000 models)
Drive 
DLT-2000 15Gb/30Gb (~$700)
DLT-4000 20Gb native / 40Gb compressed (Max) (~$1000)
DLT-8000 40Gb native /80Gb compressed (~$2000)

Software: 
Open source: Amanda (http://www.amanda.org/) available as an rpm. No experience configuring it, though... 

Closed source software:
e.g. CA Arcserve (have not tested this to backup linux), 
Veritas netbackup (v 3.41 does not work very well with Linux clients) ,
Legato Networker (Supposed to work great with Linux. In the process of being implemented here. Very costly.

Firstly, it may be worth checking with your department/univ. if they have a central backupp facility and if they would be happy to accomodate your backup requirements.

If you decide/have to create do backups to tapes on your own, you will have to make sure the backups gets stored in  safe place (in a data grade firesafe locally) and that the tape drive is not in a public location (If somebody gets hold of your tapes, your data is fully accessible to them (regardless of file permission and so on!!!)

Nicolas


-------------------------------------------------------------------
Nicolas Bertrand
CEH Oxford,
Mansfield Road
Oxford OX1 3SR
Tel: 01865 281 658
Fax: 01865 281 696
email: nsb at ceh.ac.uk
-------------------------------------------------------------------

>>> Milo Thurston <mith at ceh.ac.uk> 11/13/02 09:40am >>>
In message <B9F75223.741D%pkerrwall at psu.edu> you wrote:
>What type of hardware should I purchase and do you have any example scripts
>for such a task?

The cheapest and simplest backup setup is to use dump to store compressed images
of each partition onto a large partition on another disk. I have a couple of machines
containing two disks for this purpose, but it's also possible to dump across the network
using ssh. I can supply a couple of simple scripts to do this, if anyone wants.
This system is cheap and simple for one machine, but does not protect against some sorts
of data loss (theft and fire, for example).
For more elaborate and robust solutions involving tapes, Nicolas will hopefully answer...
Milo.

--
Dr. Milo Thurston, CEH Oxford, Mansfield Road, Oxford, OX1 3SR.
'phone 01865 281658,  fax 01865 281696.
http://www.bioinf.ceh.ac.uk/lab/ 
_______________________________________________
Bio-linux mailing list
Bio-linux at ivsun01.nerc-oxford.ac.uk 
http://www.bioinf.ceh.ac.uk/mailman/listinfo/bio-linux

-------------- next part --------------
A non-text attachment was scrubbed...
Name: Nicolas Bertrand.vcf
Type: text/x-vcard
Size: 251 bytes
Desc: not available
URL: <http://www.bioinformatics.org/pipermail/bio-linux-list/attachments/20021113/4c4c9eca/attachment.vcf>

From pkerrwall at psu.edu  Fri Nov 15 11:19:59 2002
From: pkerrwall at psu.edu (Kerr Wall)
Date: Fri, 15 Nov 2002 11:19:59 -0500
Subject: [Bio-linux] Re: PSU bio-linux clone
In-Reply-To: <sdd4dd27.027@wpo.nerc.ac.uk>
Message-ID: <B9FA88DF.7471%pkerrwall@psu.edu>

On 11/15/02 6:40 AM, "Dawn Field" <dfield at ceh.ac.uk> wrote:

> Hi Jim,
> 
> Last I saw Kerr wrote the machine was fine except for the ssh issue...
> 
> Sure, we'd like a list of the problems because we can add them to the FAQ.
> Milo and Nic are gone until Monday - I believe they offered to buy a machine
> the same as yours and re-image so that you wouldn't have hardware detection
> problems, but I don't know the final status of buying that exact machine in
> the UK - plus Nic was playing around with a similar machine in the lab
> yesterday, but I'm not sure what the verdict was - issues with the
> network/graphics card?
You're right, I did say that.  Anyway, I have fixed the problem.  It was an
error in my narc.conf file.  I don't have my server log right now, but I
think I have a pretty good of what I did and didn't do.  When I rebooted,
the os found new hardware, specifically, an extra ethernet adapter (card?).
I never had the correct ethernet slot (eth0) configured correctly in the
narc.conf file.  I also had to run 'sudo redhat-config-network' and
deactivate the second ethernet card (eth1) making sure that everything in
the firewall file was pointing to eth0.  I also changed the server name to
depcla18 which is the entry in our DNS server.  I now have all services
running on the machine which is visible fro the outside world.

> We really appreciate your using the machine becuase you are so development
> oriented.  One our end, Nic Bertrand, our bioinformatics office has also
> customized a clone in his office to do a wide variety of things (run postfix
> instead of sendmail for example), Joe, our new EGTDC Data Manager
> (bioinformatician) is starting to develop on the machine (mysql, postgres and
> transcriptomics software) and I'm doing apache/Perl/Perl-cgi on another clone,
> so we hope to learn a lot more about how to make these machines as
> user-friendly as possible.
I really appreciate being able to clone your machine.  This has really
helped me out tremendously and I hopefully can help the install process go
smother for others.  Remember that my sys-admin skills really suck and
someimes I take 2 steps backwards when I edit one of these config files.
Although, I am starting to get a little bit better understanding of the
system through this install.

>>>> Jim Leebens-Mack <jhl10 at email.psu.edu> 11/14/02 21:31:43 >>>
> Hi Milo and Dawn,
> 
> Kerr is still having system admin. problems with our new Dell server.  Can
> we set up a call for Monday morning to go over the problems we are having
> and come up with a plan for working through these?
The main thing I need now is for someone to ssh and check out my files,
specifically the graphics.  I'm still scared to run Xconfigurator again
(that's what started the mess last Friday).

> Could you please send out a list of the admin. problems you are having
> before Monday morning?  With the list of problems in hand, I am optimistic
> that we can work through everything and have a stable server back on-line
> early next week.  This really must get top priority.
I would still like to call sometime on Monday and Tuesday to talk to guys
about the system.  I would also like to talk to about what I have going on
with the analysis pipeline.  Let us know when a good time to call next week?

Kerr 



From pkerrwall at psu.edu  Mon Nov 18 10:23:35 2002
From: pkerrwall at psu.edu (Kerr Wall)
Date: Mon, 18 Nov 2002 10:23:35 -0500
Subject: [Bio-linux] Re: Disk space on genomics
In-Reply-To: <200211180953.gAI9rX3K029213@bioinf1.nox.ac.uk>
Message-ID: <B9FE7027.74A6%pkerrwall@psu.edu>

On 11/18/02 4:53 AM, "Milo Thurston" <mith at ceh.ac.uk> wrote:

> In message <B9FBEE7B.7498%pkerrwall at psu.edu> you wrote:
>> Can we move /var/lib/mysql to another partition that is larger?
> 
> Yes. Do the following (as root, or using sudo):
> 
> mkdir /home/db/mysql
> cd /var/lib/mysql
> tar cfp - . | (cd /home/db/mysql; tar xvf -)
> vi /etc/my.cnf
> 
> Now, edit the entry for "datadir" so that it points to
> /home/db/mysql
> 
> /sbin/service mysqld restart
> rm -rf /var/lib/mysql # optional!
> 
Can you do this on genomics so that I don't have to move all of my work
immediately.  

Here is the output from df:
Filesystem           1k-blocks      Used Available Use% Mounted on
/dev/hda6              4032092    226328   3600940   6% /
/dev/hda1                23302     16615      5484  76% /boot
/dev/hda7              1011928    945476     15048  99% /var
none                    515332         0    515332   0% /dev/shm
/dev/hda12            26794328   3530464  21902760  14% /usr
/dev/hdb1             65520588  27725284  34467040  45% /usr/db
/dev/hdb2             10080520    576292   8992160   7% /usr/local
/dev/hdd1             80632188  31552712  44983472  42% /home
/dev/hdd2             34746432  32211524    769876  98%
/usr/local/systemimager
/dev/hda5              6048320    318468   5422612   6% /usr/software

I can't add anything to mysql right now since /var is full.

Kerr

On 11/18/02 4:53 AM, "Milo Thurston" <mith at ceh.ac.uk> wrote:

> In message <B9FBEE7B.7498%pkerrwall at psu.edu> you wrote:
>> Can we move /var/lib/mysql to another partition that is larger?
> 
> Yes. Do the following (as root, or using sudo):
> 
> mkdir /home/db/mysql
> cd /var/lib/mysql
> tar cfp - . | (cd /home/db/mysql; tar xvf -)
> vi /etc/my.cnf
> 
> Now, edit the entry for "datadir" so that it points to
> /home/db/mysql
> 
> /sbin/service mysqld restart
> rm -rf /var/lib/mysql # optional!
> 
> That should get everything working in the way that you want.
> 'bye,
> Milo.
> 
> --
> Dr. Milo Thurston, CEH Oxford, Mansfield Road, Oxford, OX1 3SR.
> 'phone 01865 281658,  fax 01865 281696.
> http://www.bioinf.ceh.ac.uk/lab/
> 



From nsb at ceh.ac.uk  Tue Nov 19 14:44:16 2002
From: nsb at ceh.ac.uk (Nicolas Bertrand)
Date: Tue, 19 Nov 2002 19:44:16 +0000
Subject: [Bio-linux] GNUmail as an alternative email client on bio-Linux?
Message-ID: <49EA15B2-FBF7-11D6-9215-0030657C983C@ceh.ac.uk>

I know there is alreday an extensive range of email clients bundled with 
bio-linux (kmail, mutt (Command-line), mozilla mail, evolution...) 
However, GNUMail may be worthwhile addition to that already extensive 
collection?

GNUmail is a clone of the ancestor of the (i.m.o excellent) email client 
that is built-in MacOSX (http://www.apple.com/macosx/jaguar/mail.html).

Linux Focus Article: 
http://www.linuxfocus.org/English/May2002/article241.shtml

Project Homepage: http://www.collaboration-world.com/gnumail/



From pkerrwall at psu.edu  Tue Nov 26 10:49:30 2002
From: pkerrwall at psu.edu (Kerr Wall)
Date: Tue, 26 Nov 2002 10:49:30 -0500
Subject: [Bio-linux] Network question
Message-ID: <BA09023A.75E3%pkerrwall@psu.edu>

Milo,

Whenerver I use '/sbin/service network restart' it starts eth0 and eth1
which was causing me problems when I first installed a couple of weeks ago.
I uses 'sudo redhat-network-config' to make the following network changes:
    
In the devices tab, I deactivate eth1 and edit eth0 to statically set the
following IP info (Address: 192.168.1.28, Subnet Mask: 255.255.255.0,
Default Gateway Address: 192.168.1.1, these values are from the biology
system admin who is a windows sys admin, so he's no help in this).  In the
Hosts tab, I have 2 entries - IP: 127.0.0.1, Name: localhost.localdomain,
Aliases: localhost & IP: 192.168.1.28, Name: depcla18, Aliases: fgpclone.
In the DNS tab, I set hostname: depcla18, primary DNS (192.168.1.3 &
192.168.1.11), and the DNS search path to bio.psu.edu & psu.edu

>From a previous email:
>>boot up process, I get the following response >repeatedly while the boot
>>process continues:
>>Martian source 128.118.180.255 from 128.118.180.129 on >ev eth0

>This is not a serious problem. It refers to an unusual type of packet that is
>being detected on the eth0 interface. When your machine is working, I can ssh
>in and fix things so you don't see this, if you wish.

How do I turn off eth1? Would this fix this problem? And when I execute the
network restart command, how do I set eth1 not to activate?

I just got things working (not sure how), but if you could help me address
the other questions I would be happy.  You can also ssh to the machine using
depcla18.bio.psu.edu (I just ssh'd from genomics).  I'm still having a video
card problem as well.  Let me know what you think.

Thanks,

Kerr

>Kerr,
>The command that you need is:

>/sbin/service network restart

>You only need to issue the iptables restart command if you've altered
>/etc/narc/narc.conf or narc-custom.conf.

>Having done that, ifconfig should show the correct values.
>If it doesn't work, please let me know.
>Thanks,
>Milo.

>>Hi Milo,
>>
>>The biology network at penn State just added a firewall and I am now trying to
>>configure our servers for the new setup.  I was able to use
>>redhat-config-network to change the ip address, dns, etc.  Is there a command
>>that I can use to make the changes take place or do I have to reboot?  I have
>>another machine here that I was forced to reboot to get the changes to take
>>effect (ie, the ip address wouldn't work and I couldn't browse the internet
>>unless I rebooted, but once I rebooted everything worked fine).
>>
>>I've got some jobs working and I don't want to disturb them.  Also, it would
>>be nice to know commands such as these.  I tried the 'sudo
>>/etc/init.d/iptables >restart', but that didn't seem to work.
>>
>>Thanks,
>>
>>Kerr