CentOS 5 :: Automated Install Of 20 Identical Boxes
Feb 12, 2010
What's the consensus on the least hassle-prone method to do a bare metal install on multiple machines? I've just been handed a network with 20 identical servers (they're ~4 year old HP quad-Opteron machines with identical hardware configs). They're currently running a mishmash of stuff and I've been told I can re-purpose them. I'd like to create a generic kickstart and pave over them all with CentOS 5.4 with the only difference being the IP address for each machine. I have physical access to the machines and already know the MAC address for each one. Sadly, they are racked up and have no floppy or optical drive, but they DO have an exposed USB port on the front panel.
I'm trying to install CentOS on one of those BSI portable boxes. When installing, it gives this error:"Errno 30 Read-only file system." Which is weird because it's obviously not read only, and doesn't fail right away. It's an IDE optical drive, and a SATA hard drive. I thought it was the SATA optical drive to IDE HDD install issue that has been seen before, but not the case here. The disc is in known-working condition.
At work, I have a laptop configured with CentOS 5.5 that we will be using to run latency tests on two different mobile networks. To do so, we have two Novatel USB HSPA+ wireless modems connected, and configured using wvdial. The problem we are having right now is, wvdial is configured using the path to the modem to connect (/dev/ttyUSB0 for Network#1 and ttyUSB5 for Network#2), which only works if we insert the modems in the right order. I've tried alot of udev rules and what not to try and hard code the device names to something like /dev/modem_network1 and /dev/modem_network2, but that doesn't seem to work.
So at this point I'm trying to take a step back and a fresh look at the problem. Can anyone give me any information regarding where I should begin to get this done?
I installed webmin on many servers under my control and decided to automate the job. Here is a shell script to automatically add WebMin repo to yum and install it. You can use "nano -w webmin_install.sh" to create a script somewhere, copy there the source, save, then allow execution using "chmod u+x webmin_install.sh" and then use "./webmin_install.sh" to execute it.
I am trying to use two CentOS 5.5 boxes to serve up email. I wish to loadbalance HTTP, IMAP, POP3, and SMTP between the two servers. The servers are not in a cluster (the email software does not require it). What filesystem could be used to distribute/replicate the files between the two servers? I have tried NFS, GFS, and GlusterFS. All of which work but result in a latency that is unacceptable to the end users (webmail clients).
My X-windows *server* is a Cygwin box. I installed the entire X11 packages. on the centos machine, I installed xterm and emacs.xterm comes up fine and the characters in the window are fine.with emacs, the characters in the windows are empty boxes.Do I need to install some things on the centos end to get the characters to render?
In CentOS 5.4 (Final), emacs is displaying characters as little boxes. see attached screen shot. I searched the Web and found others have this same problem with emacs in CentOS 5.4 but have not found any solution. I installed emacs in a base CentOS Amazon EC2 instance as follows:
My first time installing centos server and adding it to xp home network. I am new to setting up my own network in general. I have a home windows wired ethernet xp network (simple linxs router) with 2 xp boxes and one centos5 box all connected to the same router. I just set up the centos box running tomcat on port 8080. I need to do two very basic things at this point, but am not sure what I need to do:
What do I need to do so I can: 1. Connect in firefox on one of my xp boxes and call the tomcat server running on my centos box? 2. Be able to ftp to centos box from either of my xp boxes?
I'm having a bizarre problem with FTP transfers from my CentOS 5.3 boxes, All of my CentOS boxes appear to be limited at ~10kb/s when connecting to any FTP server outside of our network. We have no network infrastructure which would cause this, however when I use a windows based machine it is able to connect fine.
I took my laptop running Vista, and rebooted it onto a CentOS 5.3 LiveCD, and same issue.
I have checked our iptables, cleared the rules, removed the kernel modules, checked CBQ/TC it looks like nothing is limiting this. Also, other protocols work fine. Both the windows and linux boxes used passive transfers for it.
Is there anything other than iptables/CBQ which would be on a out of the box install of CentOS I should check? Is there any kernel settings for NAT connections which may cause this?
I am new to Linux. When I installed some applications and something through package manager,Now I get two added options in boot loader with identical names like failsafe and normal. what is that?multiple kernel?
I am trying to set up an automated installation system for hardware testing and was wondering if anyone on here can point me in the right direction.I would like it to be compatable with a wide range of OS distros including Ubuntu, Red Hat, CentOS, Suse, Windows, Dos, FreeBSD, etc. If some custom scrypting is required thats fine, but flexibility is key.I am looking into Jumstart and Kickstart right now, but I am not sure if that is what I need.
I tried remastersys, but has to install graphical stuff for the ubiquity. I don't need an installer, nor any gui packages. Is there another good automated program or script that works great for making compact text/console only installs?
i have a windows xp desktop machine that i've been using as a fileserver throguh sharing my 4 external usb drives attached to it. I want to replace this sytem with ubuntu (unless theres a different alternative for waht i want to do?). the problem is its not a standard install i can just do.
- first, the desktop has no keyboard mouse or monitor connected and i generally used vnc to control it.
- the computer right now is heavily infected with some kind of virus. disabld all my anti virus software, wont allow me to use new ones, etc.
my only option really is to wipe it clean and this time not xp, i want ubuntu. is there a way to configure an ubuntu install (or find a preconfigured install) to basically boot automatically from cd/usb drive when i restart my computer, do the install with no user intervention, and upon finalization install a vnc server on the install so i can get back in.
please remember this is my first time wiht ubuntu and while i know more than your average user about computers, id ratehr not muck around. i jsut want to be able to download a standard install, put a file on the installation cd or usb drive and let it install itself.another note, 4 of my external drives connected via usb are ntfs, and 1 is hfs. can ubuntu read/write to all of these systems? if not out of the box, can i later enable these filesystems somehow?
I'm setting up a large number of virtual machines, each with a basic set of about 30 applications. Obviously, I don't want to do everything manually so I'm looking for ways to automate this process. Multiple distributions (Ubuntu, Red Hat Ent., FreeBSD, CentOS, etc.) will be used, meaning I will have a few Ubuntu installs with the basic set of 30 applications, and multiple Red Hat installs with the same set, etc. So, I'm looking for advice on automating as much of this as I can, even if it means a new form of automation for each distro.
I have discovered that I have two partitions, on separate hdd's with identical UUID's, and the system switches back and forth erratically on restarts between the two different partitions, giving me the current /home or the /home of two months ago when I did the upgrade.
How do I boot with two identical drives? I have two identical Western Digital WD6400AAKS SATA-II 7200 rpm 640 GB hard drives. One of the drives is an internal drive on one machine. The second is available with a mobile drive bay. I can't boot the system with both drives installed. The kernel boots fine but then halts when handing off to init. The error message is:
Warning: unable to open an initial console ... Kernel panic - not syncing: No init found. Try passing init= option to kernel.
I can place the mobile drive in a different system that does not have the same drive installed and that system will boot fine from either drive. I'm reasonably certain the problem is caused by the two drives being so identical. I suspect the problem is the kernel cannot distinguish the difference. The model numbers are identical. No, I don't have other SATA drives to use. I am not using raid. The BIOS is not configured to use raid. How do I boot with two identical drives?
i have an sql table with 2 columns i run a script that randomly selects a word from the table in column 1. the word is displayed on the screen and I guess what it means i concatenate the randomly selected word and the answer the script looks for a match in mysql if it finds a match it says "Good job!" if there is no match it will say "not correct". However when i get it right it says not correct even though when i echo the variables they look exactly the same. the script below:
#!/bin/bash var=$(mysql translator -u root --password=*-N<<EOF SELECT word FROM tagalog ORDER BY RAND() LIMIT 1 EOF )
I have two identical 73 GB Scsi ulta320 scsi drives, Fedora web server is install on one drive with all web files and etc. I wish to make an exact clone of the drive that will boot and run everything as the current drive does now. Is there a download of a program I could download or purchase that would boot and make an exact clone to do the above.
I have two identical hard drives; same make, same manufacturer, same model, and same capacity, which I'm trying to run in a RAID1 mirroring scheme. The problem: configuration files for md arrays only lets me use device names, such as /dev/sdb and /dev/sdc. To keep these the same (and in the same order) when I boot I wanted to write a udev rule for them. Unfortunately, I have no way to differentiate between these two drives, as they seem to be identical. Normal methods of differentiating by size or model name wont work. I think I can use UUIDs; but I neither know how to get the UUID of a device/partition, nor do I know how to use it (if it is possible) in a udev rule.
Solution:
run udevadm info --query=all --path=/sys/block/sdb # or whatever block dev
Look for and use "ID_SERIAL_SHORT" which is unique even for identically manufactured disks. Write a udev rule based on this property.
I'm running Edubuntu 10.10 64 bit on an LTSP box here at work. Two of them, in fact. In each instance, we have a Ricoh MP2500 network printer.
Scenario 1 - I print, it works. Scenario 2 - I print, the printer heats up, the wheels spin, yet it never actually prints a page. Each box is from the same Edubuntu install CD. Each box is set up as identically as possible with very minimal differences (aside from software set, as one is for Middle School and the other for High School) Each scenario has an identical printer. Each scenario has the same printer driver.
I have two identical laptops. One has an installed Ubuntu and parts that I want to use. The box itself is all beat up. The other box is newer but has a stale edition of linux. (Stale software means "seldom used.") I thought to be clever I would pull drive-A from box-A and install it into box-B. Likewise, I would install drive-B into box-A. This will leave my clunky software on my battered box and my newer software on the newer box.
Mechanically it works (doh!). However, neither box will see the network. When I look at the logs, I find a "rename wlan0 to wlan1" entry among others. If I put the drives back into their original box, all works correctly.
What could be going on that I cannot move the drives and have things just boot and run?
I thought that system start would detect the installed hardware, load the required drivers and all is right with the world. The "rename" log entry suggests that the old hardware details are somehow in the way of the new hardware discovery and configuration.
Is there some command I need to use or utility that I ought to run that says, "rediscover my hardware" or similar?
I already know how to use clonezilla and other ways to duplicate a drive contents to a second drive. (Note to reader: laptop drive to usb drive clonezilla takes quite some amount of time.)
Another reason that this is important lies in the ability to move a drive from in-use but failed hardware to stand-by working hardware in a fail soft recovery situation. I know that win-doze knows about the installed hardware and demands a re-install or "repair" to the alternate box... but this is linux not win-doze.
I have two servers (A and B) which are identical. My idea was to have a software RAID 1 with both of them running and if server A craps out I wanted to swap the hard drives from server B into server A. When I tried to test this idea the network driver doesn't seem to want to cooperate. ifconfig gives nothing and I can't figure out why. From googling I've read that a network card has a unique ID on it (I'm assuming their talking about the MAC address) and is used in some config files which is why swapping hard drives gives two different MAC addresses and confuses the system. If that is so would anyone know exactly what config files the mac address is stored/used in? That way I can make a backup of those files and swap those out if I ever need to swap out the hard drives.
I have two identical servers, one has RHEL 5 and Zimbra installed and the other is currently not really doing anything. Both have hardware RAID (Adaptec) set to RAID10, identical hard drives, etc. The RHEL/Zimbra machine is set up with LVM2. Is it possible for me to hook them up on the secondary NICs and boot the second machine with Knoppix or something else, and easily tell it to duplicate the first machine onto the second, down to the last bit, or do I need to make all the partitions beforehand and dd each one separately?
I am trying to do a comparison of two folders, let's call them dir1 and dir2 and remove any items that have the same file name in both folders from dir2.
The goal is to make connection calls (ssh, ping, ...) possible from one LAN (LAN-1) to a number of (at the moment two) separate smaller LANs.These smaller LANs (LAN-2a, LAN-2b, ...) have exact same specifications (same IP range, same number of nodes, ...)!The idea is to use a Fedora box (release 14 with 2.6.35.6-45.fc14.i686) and implement an appropriate iptables routing/forwarding.The Fedora box has three network interfaces:
- eth0 (aaa.bbb.ccc.m) on LAN-1 (aaa.bbb.ccc.0/24) - eth1 (ddd.eee.fff.n) on LAN-2a (ddd.eee.fff.0/27) - eth2 (ddd.eee.fff.p) on LAN-2b (ddd.eee.fff.0/27)
i have a problem in finding block of identical strings...i solved the problem in finding consecutive identical words and now i want to expand the code in order to find and remove consecutive identical block of strings... for example the awk code removing consecutive identical word is:
