Slackware :: Setup Software RAID1 On A Running System?
May 23, 2011
I have a nas / server running Slackware64-13.37 with two raid arrays, one with 3x500GB disks and the other with 2x2TB disks. I also have 1x80GB hdd which contains the os. I found a spare 250GB hdd from my closet and want to set up raid1 array from os hdd and that spare drive.I know that I can't make use of the extra 170GB of space but I don't care as long as I have working installation on the other disk if that 80GB hdd breaks. How can I set up that kind of raid1 without losing any data?
So I didn't notice when I setup my CentOS 5.5 server that I left / as RAID 0 on md1. All the rest are RAID 1. Is there a way I can modify the array to RAID 1 without a risk of data loss? I'm glad I caught this before I setup any other services. I've only setup smb so far...
[root@ftpserver ~]# df -h Filesystem Size Used Avail Use% Mounted on /dev/md1 16G 3.0G 13G 20% /
I recently set up an old desktop computer on my home network for use as a file server. I installed Ubuntu 10.04 Server on it and I'm slowly learning how to make it do what I want (I'm used to the GUI, so this is a bit of a jump for me).My plans are to put two 1.5TB hard drives in the computer (there's already a 160GB with the OS installed) and put them in a RAID1 configuration. This, as I understand it, will write all data to both drives, creating two identical drives. I'll only have 1.5TB total space for backup, but if one of the hard drives die, I can replace it without losing any data.
Hopefully I understand all that correctly. Now, my questions are how to actually set a RAID configuration up. I don't have a controller or anything, assuming I could do it off the two SATA ports in the motherboard (the 160GB is IDE) in software RAID.Once I have the two hard drives set up in RAID1, how do I allow the house to access the file storage?We're all on Windows 7 computers.If one of the hard drives fail, how will I know? Can I set it up so I receive some sort of notification? Also, how does the replacement process work; can I just pop the broken drive out and put a similar one in? Do I need to readjust any settings after replacing one?
I decided to use my old computer for torrent downloads and perhaps for mirroring the slack -current. However, the computer is rather weak: Celeron 1,2Ghz, 256 MB RAM. So, which slack release is better to use? I aim to use the jfs file system, but I'm not sure it is good choice.
I searched, and unfortunately all that did was raise my confusion level on this. "Grub"? "mdadm"? "fake"?, S/Wraid? Disk utility? So many options, so little understanding! Relevant stuff (Mostly reported by "Disk utility"):
- Lucid Lynx. PATA host adaptor -> IDE controller -> Maxtor 164GB H/D [I *think* this has Windoze on it, but can't remember!] There's also a CD & DVD drive on this IDE bus. PATA host adaptor -> SATA controller -> Seagate 500GB H.D. - The Ubuntu boot drive, currently a 250GB ext4 partition, 3GB swap and 250GB "unused".Peripiheral devices -> Firewire 400 -> 2 x Samsung 500GB H/D's. These were "stolen" from my Mac Book and are currently a RAID1 Apple array. Everything they contained is safely backed up, and these can be considered as "new" drives awaiting formatting. [It's actually a Buffalo Drivestation Duo, but their site was even more confusing than here.
everything is working wonderfully, but I'd like to use the 2 F/W drives in a RAID1 array - So, eventually to the question: How do I tell Ubuntu to use these drives as a RAID array? It seems I can format and partition etc from disk utility. Do I then use mdadm for configuration? Any other recommendations?
I'm trying to take my current windows7 (x64) setup and make it into a dual-boot setup with slackware 13.0 (x32) and am hitting a brickwall. I guess I would like to know if I can salvage the situation without formating my windows drive. I fully intend to, but I had hoped to not do it this month. I'd like to get dual boot going so I can master it and acquire all the files/drivers I'll want and then format and do it 100% right next time.
Currently I have windows setup on a 3x 500gb Raid0 onboard array which itself has been flawless. After I shrunk my current partition size to give me about 25gb of free space I proceeded to setup Linux, and I was unable to perform the cfdisk portion for partitioning the array for Linux.
I tried using cfdisk: /dev/hdx (a1-a3,b1-b3) /dev/sdx (a1-a3,b1-b3)
I even attempted to locate with: cat proc/partition and tried using cfdisk on every device it located. It always said it was either an unknown partition table and should I start at zero, or bad partition. I was of course too concerned over my windows setup (which has almost 1tb of stuff I have not backed up) to go any further into the unknown.
PS: I have used Slackware before, back at 8.1, 10, and even a very brief interlude at 12.0, this is my first attempt on this computer however with 13.0, it is currently running fine on my laptop and my PS3 (Laptop is even dual booting 7/Slack13. Albeit without raid)
Phenom 9850 M2N-SLI Deluxe (Nvidia AM2) 4x1gb of 1066 kingston hyperX 3x500gb WD Caviar Black Sata2 3.0
I am trying to get PHP 5.3.4 into my webserver, running Slack 13.1 on a 32bit system. Basically, it's a success, but no matter what options I try, phar will not be installed
Code: Generating phar.php Generating phar.phar PEAR package PHP_Archive not installed: generated phar will require PHP's phar extension be enabled. clicommand.inc directorygraphiterator.inc
I'm running a software RAID1 on 3xSATA drives. I'd like to make one a spare. How do I do this? Is there a way so that once I do when I have a drive failure the spare will automatically mount and be made a mirror?
I have been playing with chroot to build packages.I am wondering about the limits or caveats of using chroot to mix operating systems. For example, what should I beware if I use 12.2 and chroot a 13.1 system? Or if I use 13.1 and chroot a 12.2 system? After all, the chroot does inherit some attributes of the parent system.There is the noticeable difference of kernel versions. For that one reason I'm thinking that mixing system versions might be frustrating.
My focus is building packages, not running an alternate operating system or a service such as ssh. I realize that building packages as root in a chroot potentially compromises security but that is not my focus.I am aware that a chroot is not a true virtual system.A virtual machine is an option, but in this thread I'm curious about the limitations of chroot. I find building packages in my virtual machines much slower than actual hardware, even with a dual core. Hence my interest in chroot
I was experimenting in virtual machines this evening with software RAID.I found Slackware's method straightforward and intelligible. The README is excellent. I found software RAID setup in Scientific Linux (RH Enterprise clone) and Debian convoluted and confusing.My only question regarding software RAID in Slackware is how the system notifies you about the failure of one component. I pulled one of the virtual disks offline to deliberately fail the array but I wasn't notified when rebooting. I didn't look too deeply into log files to be honest. Can it be set to automatically notify the admin about a failure?
My old workhorse computer's motherboard died yesterday, and I want to get everything off of its RAID1 array. I have a backup on an external drive but it's a few weeks old and I'd like to make sure I've got everything.
The old machine ran Slackware 12.1, and had a 2-drive IDE 250GB RAID1 array with partitions:
md0 - swap md1 - /
The new machine has Slackware 13.1, and also has a 2-drive SATA 250GB RAID1 array with partitions:
md0 - / md1 - swap md2 - /home
I put the IDE drives from the old computer into the new one. I'm not sure how to get the old array going now. I'm not sure if I should use mdadm with --assemble (since the array was already set up before) or with --create (because the array needs to be renamed so it doesn't clash with the new computer's md1). I'm thinking I should use --create and give the old md1 a new name (md3). But I'm not sure if anything bad will happen if I use --create on an array with data on it.
The old drives are the last two entries, sdc and sdd. It's odd that the order is reversed. I had (non-RAID) Windows partitions on both drives and I've mounted them and verified that the drives are on the IDE cable in the right order, and sdc is the original 1st drive of the array, and sdd is the 2nd drive of the array.
have been trying to setup a dual boot system with ubuntu and XP running side by side on my Thinkpad T41.tried it a few times and always causes the same problem. i have 40 gig HDD, on which i create a 13 gig NTFS partition and leave the rest as free space. then install XP on the NTFS partition. no problems.
then i boot from the ubuntu disk (9.10 Karmic) and install using the "use free space" option at the partition section. ubuntu installs ok, and boots fine from GRUB 2.0. BUT when i select the XP option from GRUB's list, it starts to boot XP, i get the standard XP loading screen for three seconds and then it crashes to a blue screen critical problem, and restarts the system. when i then boot from the xp cd and go into recovery mode CHKDSK will not recognise the disk, and DISKPART shows one HDD at 35 gig which it cannot access.
this means i cant run FIXBOOT and get my xp install running again. every time i do this process it produces the same problem. tried at first with xp installed on whole HDD, and reducing the xp partition size. killed XP. then tried ubuntu first and xp second - but this caused the same inaccessible disk problem - xp would not recognise the partitions and would not install. so i slipstreamed my XP install disk to SP2 hoping this would make it recognise the partitions, but no luck there. so had to format all and repartition the 13 gig NTFS for xp. installed xp again without difficulty but ubuntu install killed my xp in the same way.
I have an existing Fedora 15 system installed from scratch.I've ordered a harddrive identical to my SDA and want to add it to my existing system as a RAID1 setup.I've googled around and cannot find recent clear instructions how to accomplish this. I don't want to reinstall everything from scratch. It should be possible to create the RAID1 using the existing data disk and then mirror everything up?
I am trying to migrate my existing system with one IDE disk , tools installation already done... without loosing informations and having to install once again every things, to RAID1 (soft) with a second IDE disk I tried to do this using somme informations given on forums but i always have a kernel Panic at the end of boot What I did:
The system is going down for system halt NOW! login as: root root's password: /usr/bin/xauth: creating new authority file /root/.Xauthority
I've got a couple of commercial NAS boxes and I'm wondering if they (ReadyNas duo, DLink DNS-323) or any other NAS is suitable for having their RAIDed disks moved to a software-based NAS. To be specific, I'm a big fan of the (largely) Debian-based Ubuntu. Can the aforementioned NAS drives be migrated to Ubuntu (e.g. using the mdadm Linux command)?
Secondly, is there any commercial NAS that can be migrated over? Incidentally, here is a link to somebody who succeeded in a migration:URL...My specific scenario I'd like to prepare for, is the eventual (sudden) death of one of the NAS motherboards.
I've read many of the postings on ICH10R and grub but none seem to give me the info I need. Here's the situation: I've got an existing server on which I was running my RAID1 pair boot/root drive on an LSI based RAID chip; however there are system design issues I won't bore you with that mean I need to shift this RAID pair to the fakeraid (which happens to most reliably come up sda, etc). So far I've been able to configure the fakeraid pair as 'Adaptec' and build the RAID1 mirror with new drives; it shows up just fine in the BIOS where I want it.
Using a pre-prepared 'rescue' disk with lots of space, I dd'd the partitions from the old RAID device; then I rewired things, rebooted, fired up dmraid -ay and got the /dev/mapper/ddf1_SYS device. Using cfdisk, I set up three extended partitions to match the ones on the old RAID; mounted them; loopback mounted the images of the old partitions; then used rsync -aHAX to dup the system and home to the new RAID1 partitions. I then edited the /etc/fstab to change the UUID's; likewise the grub/menu.list (This is an older system that does not have the horror that is grub2 installed) I've taken a look at the existing initrd and believe it is all set up to deal with dmraid at boot. So that leaves only the grub install. Paranoid that I am, I tried to deal with this:
dmraid -ay mount /dev/mapper/ddf1_SYS5 /newsys cd /newsys
[code]....
and I get messages about 'does not have any corresponding BIOS drive'. I tried editing grub/device.conf, tried --recheck and any thing else I could think of, to no avail. I have not tried dd'ing an mbr to sector 0 yet as I am not really sure whether that will kill info set up by the fakeraid in the BIOS. I might also add that the two constituent drives show up as /dev/sda and /dev/sdb and trying to use either of those directly results in the same error messages from grub. Obviously this sort of thing is in the category of 'kids don't try this at home', but I have more than once manually put a unix disk together one file at a time, so much of the magic is not new to me.
/dev/md0 (made from sda1 and sdb1) RAID1 /boot partition /dev/md1 (made from sda2, sdb2, and sdc2) RAID5 / partition
Earlier on I had some trouble with my sda drive, it dropped itself from both arrays, screwing up the mirroring of my two raid partitions participating in the /boot partition. I eventually got everything sorted out and back in sync. (I also have grub installed to MBR on both sda and sdb). Things are working fine regarding that, but since then I've had this issue:
During boot up, I'll get an error message that it could not mount my /boot partition (when fstab is set to either /dev/md0 or the UUID). It claims c9ab814c-47ea-492d-a3be-1eaa88d53477 does not exist!
My fstab:
Code:
[mark@mark-box ~]$ cat /etc/fstab # # /etc/fstab # Created by anaconda on Wed Jan 20 16:34:41 2010
[code]....
As far as I know, it isn't neccessary for /boot to be mounted always, correct? Although, as I understand, I need to have it mounted whenever making kernel changes correct?
I just made a complete reinstall of my fileserver. After the installation the system cannot autostart my Raid1. For some reason it seams that only one of the identical disks are found by mdadm"disk utility" states "not running, partially assembled"If i immediately stop the raid in the disk utility I can restart it and mount it.some diagnosticsmdadm.confQuote:
# by default, scan all partitions (/proc/partitions) for MD superblocks. # alternatively, specify devices to scan, using wildcards if desired. DEVICE partitions
I have an Intel server, which has it's two SATA HDD's in "Intel Embedded Server RAID Technology 5.4" RAID1 volume. How to proceed with a system image in case two of those SATA HDD's fail at the same time? Should one take the first HDD of RAID1 volume, connect it to another machine and execute:
Code:
# ddrescue /dev/sda1 /media/External/image_of_first_hdd /media/External/log_of_first_hdd * HDD from the problematic RAID1 volume would be recognised as /dev/sda1 behind new machine * /media/External/ is a mount point for large external HDD in the new machine * log_of_first_hdd would be the log file
..and then take the second HDD to another machine and execute:
No, not Arnold S.I'm talking about the one from slackbuilds.org. It's not just another terminal emulator. It's got a cool feature of broadcasting the input to all (or selected windows). If it's a well known and used feature, than I'm sorry. I read about it a couple of days ago and soon it proved a big time-saver.
As all my systems are almost identical (selection of software, /home/ directory structure, program configuration, etc) I used terminator to set up my new shiny Slackware64 13.1 systems.The 'broadcast' function of terminator mirrors your input to other terminator windows (ssh sessions to other machines). That way I didn't have to configure it three times on three computers. I know that one can write a script to automate most of things, but sometimes there are interactive steps which would not work in a script. It was especially useful in:
1. using sbopkg (3 simultaneous instances doing exactly the same thing on different systems ) 2. general configuration 3. setting up bookmarks in emacs 4. Creating some scripts that aren't there by default (eg. /etc/rc.d/rc.local_shutdown)
I can imagine how useful it can be for administering a number of machines.
My system decided to crash on me, hard. It was humming along happily for about 2 months and now doesn't boot. If I boot from hard-disk, I get grub. Launching the first kernel choice hangs. I thought maybe the install was corrupt, so I booted from usb install disk. The usb hdd didn't boot; something about an error trying to access /dev/sda . Unplugging the internal disk and plugging in the usb install disk does result in the system booting. Plugging in the internal disk in a running system usb-booted system does not result in the system detecting the disk.
How do I know if the disk is physically broken? This seems unlikely since it does manage to launch grub consistently. Or is this still possible? How can I try to mount whatever is left? The usb install disk doesn't even list the /dev/sd*. Any pointers on how to reformat the drive if it's not being mounted?
I yesterday installed Slackware and I am quite impressed. I have been using Ubuntu for a year and decided to move on. My wireless worked fine in Ubuntu but it seems that it needs some configuration for Slackware.Note that my home network WiFi has no password or whatever. I still can't find any networks with my wireless even after the configuration. I fire up wpa_gui and it can't spot anything..
I only see two options in KDE. System settings which shows nothing relating to printing and Office Org printer administration.In the latter Office org printer admin shows "generic printer"..My printer is HP 5P. and I choose the option 5P/MP.Then next step asks me to: enter a command line for the appropriate device. I don't know what the appropriate command is. Can anyone help I leave it empty and click next. It shows the 5P/MP is now listed along with generic. I make it default and send a print command resulting in "error while printing".
I am an absolutely beginner with Linux (ubuntu). I got fed up with widows and decided it was time to go. I installed (clean install from a cd) ubuntu 10.04 on my desktop. It is an hp compaq d220 MT.It looks good and feels good, well at least for a while since it keeps crashing. The screen starts to flicker, goes black, and the only way to get it up and running again is to press the on/off button until the system powers off and restart.
What would be the advantage of running a 64bit system over a 32bit system? I only have 3GB of ram but plan on kickin another 1GIG into it. But i wanna try 64bit Linux(probably slackware) on it. But first im just wondering what the advantages are people have seen who have used both 32/64bit linux. Speed? Smoothness? And also what are major disadvantages such as compatibility, configuration, etc.
I'm new to slackware and I am going to try install 13 but I would like to know how well MythTv will run on slackware 13. There are so many parts to MythTv that I thought there might be issues with depedencies not being up to date, etc.If you have it running on 13 were there any issues to work through?
apt-get install startxwould sort this out but no joy. I don't seem to actually be able to get anything using apt-getI've tried running apt-setup but the command isn't recognised.I'm sure I'm connected up to the internet (I'm using a wired connection I run dhclient eth0 and check with ifconfig and eth0 seems all happy and connected up). I'm also trying to install mpg321 or mpg123 as it would be nice to have some tunes but I get the same "package not found" error message
