I've spent the better part of an afternoon looking for a solution to a problem: backing up my installation of 10.10 as an image file to an external hard drive. My research has yielded a lot of suggestions for clonezilla, dd, and partimage/particlone, but those don't seem very appealing, due to a number of issues (can't backup live, copies free space as well, doesn't handle ext4, etc). Also why is clonezilla 150mb?
I'd like a simple solution that can clone an entire disk (used space only) to an explorable image file on a separate hard drive and be able to do it while the operating system is running on the disk. I used to use apricorn ez gig to do this on windows and it worked like a charm, but I can't seem to find a similar solution that creates and explorable .iso image file with linux. I've used superduer on osx, which is awesome and i wish there was something like that for ubuntu/linux.
We are using Ubuntu 10.04LTS server on Vmware Vsphere estate. We are using LVM and have / and /var on separate partitions.
We have been experiencing an odd issue with sda1 always being fsck'd after every reboot. We seem to have traced this to the start/kill scripts in rc0.d and rc6.d. It appears that the reason the disk is not being unmounted is because some of the scripts are never being run because they are prefixed with S. We renamed and reordered the scripts to reflect what we thought should be the correct order - i.e. not halting the system before unmounting disks. System reboots cleanly now.
Before list of rc0.d and rc6.d respectively: lrwxrwxrwx 1 root root 17 2010-11-18 15:06 K01apache2 -> ../init.d/apache2 lrwxrwxrwx 1 root root 22 2010-11-16 17:03 K01zabbix-agent -> ../init.d/zabbix-agent lrwxrwxrwx 1 root root 22 2010-11-30 08:16 K03vmware-tools -> ../init.d/vmware-tools
basically are we expected to do these changes as part of building new servers for production use? Quite happy to do this but just surprised that this is the default.
I have my Mythbuntu 9.10 environment installed on an old 160GB PATA disk and have just purchased a new 64GB SSD that I want to transfer my installation to.In the past I've just used ddrescue to clone disks, however in this case the source disk is larger than the destination disk so it won't work. I only have a few GB of actual data on the 160GB disk, so the 64GB SSD is definitely going to be enough for me.
I guess I need someway of either cloning so that only the actual data and not the partition is brought across, or possible shrinking the partition(s) first on the source drive and then using the same ddrescue method I've used in the past.Just looking for some assistance on what method is the best/most reliable?
Recently my had disk crashed. It has taken me over a day to install and setup the drive - time I can't really afford. I'd like to know how I can clone a second as a bootable copy of the main hard drive and update from time to time. Therefore if the main hd fails again, I can simple boot from the second drive.How can I do this? I have two sata drives in the box, running 10.10 desktop.
Recently my had disk crashed. It has taken me over a day to install and setup the drive - time I can't really afford.I'd like to know how I can clone a second as a bootable copy of the main hard drive and update from time to time. Therefore if the main hd fails again, I can simple boot from the second drive.
I'd like to have a bootable cd which then allows me to clone a hard drive, much like the "Ghost" application. I need to copy Fat32, NTFS, as well as the various Linux formats. I've not created a bootable cd before, so a step-by-step would be rather handy.
I use dd in its simplest form to clone a hard drive dd if=INPUT of=OUTPUT However, I read in the manpage that dd knows a blocksize parameter. I was wondering whether there is an optimal value for the blocksize parameter that will speed up the cloning procedure?
I just invested nearly 12 hours configuring a CCTV system using CentOS 5.5 Server and Zoneminder. I have it setup just the way I want it. I would like to make a clone image of the drive just in case disaster strikes (lightning strike, failed hard disk, etc). In the Windows PC world, I use a program called Ghost to make a mirror image of a hard disk. I power the computer down, run Ghost to make a block level clone of the drive, then power it back up. Can I assume that will work with CentOS without problem?
In the computer now is a 320GB SATA drive. One partition on it is swap, and the other is ext3. There is no raid setup on the drive. I have an identical 320GB drive I could use and keep it in the computer unplugged from the power and not spinning. That way if anything ever happens, I can power down, move the power and data cable to the new drive and power it back up. Granted, I will lose any new config and database changes, but it will be a lot better than starting back at square 1 and reconfiguring the entire OS and software.
I cloned F14 with Clonezilla from 80GB to 320GB hdd(both sata disks), and then resized the partitions with GParted.But I can not boot into fedora on the new/bigger disk, it stops and the display writes "Loading stage 1.5" if I remember corectly,I tried to fix it with the live cd but with no efect.
Then i found Super Grub Disk live CD, and with that i tried to use their fix, which was the same as with the Fedora live cd i tried before, again no efect.Then i played around with Super Grub, and found the option to boot GNU/Linux indirectly, and with that metod i got results, found my menu.lst file and chose the kernel i wanted and it boots into desktop.
But i would need a more permanent solution, because now i allways have to use the same procedure with Super Grub Disk CD to boot into my Fedora 14.
To make a full backup I run a live Knoppix DVD and clone the computer's HDD to an external HDD using the dd command. Is there a possible problem with the source being copied onto bad sectors on the destination disk? If so is there a way to prevent this from happening? A typical dd command I use looks like: dd if=/dev/sda of=/dev/sdb bs=4096 conv=notrunc,noerror. Is this the recommended command for cloning to a disk of equal size?
I have just installed 10.04 as an upgrade from 9.10.
I am having difficulty because none of the USB drives I am using are unmounting as they should be. normally speaking when erasing data from my USB key when you come to unmounting it it will ask if you would like to send this data to the recycle bin thus giving you back the extra storage space on the device. 10.04 isnt allowing me to send these items to the recycle bin ad doesn't give the option to do this.
10.10 MMeerkat - When I use a USB-stick or external hard drive, I don't get the "unmount" option but instead it says "safely remove drive".When I click this, Nautilus immediately crashes and is gone.Is this a known bug in 10.10?
Very frustrated. I have used Unix for ages so I understood the SysV startup stuff. But I have not had a lot of luck with Upstart. The other day I noticed that on every reboot my disks are getting fsck'd. I just recently put an ext2 on /tmp so this takes a while (the ext4 drives just rip through their journals).
The problem is no one is unmounting them on a KDE restart (4.X). I started out looking at /etc/init.d/umountfs and putting some logging in there. It never runs. This is despite that /etc/init has an upstart job that is supposed to run all the runlevel stuff.
I also tried to log some info in /etc/init/mountall-shell.conf which looks like it tries to do a umount -a on shutdown (which is probably not a good idea; you need to unmount in a particular order). That doesn't seem to happen either. I am not even sure how to troubleshoot this further. I suppose I need to see if the reboot( command has the same problem. Or if I shut down kdm first if it goes away.
I get the following error each time I try to safely remove the USB stick (by right-clicking on the USB stick icon on the desktop and choosing "Remove safely") code...
It used to work without errors when unmounting the USB from Nautilus window, but now, I get the same error when unmounting from there as well. Is it safe to assume that the stick can be removed even though the error message is displayed? What might be the cause? Also, I do not have any Nautilus-windows opened nor any applications that are locking any files on the USB stick while removing it.
I want to move my / partition to the end of my drive (sda). To do this with gparted, I have to unmount it, but I'm not comfortable with the idea of unmounting root partition... Should I do it from a live cd? More important : is the operation safe?
I have an external hard drive (1TB MyBook) mounted via fstab by UUID to a directory. When copying/writing/reading a lot of files from it, it randomly unmounts and remounts as a different device.
It'll start as /dev/sdb1, I'll set a lot of files to copy to it and then it'll unmount and re-acknowledge itself as /dev/sdc1, the file copying process will crash and the current directory in terminal will display an I/O error. Running mount -a remounts it back to the directory specified in fstab as /dev/sdc1 and the loop continues. If it's just idling, there's no issue, only does this under load.
- Nautilus open in browser mode (the default)?- I click the eject symbol to unmount a usb stick in the side pane.- The annoying red exclamation police light looking thing pops up telling me that nautilus crashed.- Not only is the usb stick unmounted, but the device file also disapears (e.g. no longer listed by fdisk).- This problem does NOT occur when I use other methods to unmount (e.g. umount on command line, or right-click on desktop icon and select eject action), only when I click the eject symbol in the nautilus side pane
been puzzling over this for a while, searched forums and net but no issues/solutions found. I have the above device plugged in and working fine. However when un-mounting - either through dolphin or device notifier, I get the report that the item is unmounted but then when removing USB cable from the Zen, it closes down and then rebuilds the entire 16GB collection - this takes quiet a bit of time and I worry that I may end up frying the Zen. Is there a fix for this in Suse or is it the zen. it doesn't happen on my work machine running windows though.
I have a volume that shows as the following when I do a df -h. How would I go about unmounting it so I can run an e2fsck on it, then remounting it? normally it mounts when the server starts, so i'm not sure how to manually do it.
Sometimes I get a problem with the basic "umount" command. I get an error message telling me that for one reason or another, the device couldn't be unmounted. Usually, it says the device is busy, when I can't see how it possibly can be. When this happens I'll use "umount -l" or "umount -f" or sometimes "eject" but I'm still not happy, because at the back of my mind I'm worried about damaging the integrity of the device's filesystem. What's the CORRECT way to deal with this problem?
I stupidly unplugget my USB-cable, which was connected to my Nokia music phone, just as if I were in Windows. What do I do? I've lost my music on the phone, or, it seems it may be there (the correct mass of data), but my phone now tells me there is no music... Can I recover this? And - what is the correct way to unplug a USB unit in Ubuntu? To make it work, and find the phone/drive - I just typed "sudo lsusb" in the terminal, and it found and opened the memory automatically... How should you unmount the USB, and maybe even how do I get my data back?
I have a server with Ubuntu 8.04.4 LTS completely patched up to date.I tried to mount a windows share using smbmount and unmounting it with smbumount. I did it as a non root user (let's say "someuser") and managed to mount the share without problems with "smbmount \\server\share xxx" where "xxx" was a directory on the user's root. After it I could see the shared files in the xxx directory, as it was supposed to be...
The problems came when trying to unmount it. Using "smbumount \\server\share xxx" didn't work giving "This utility only unmounts cifs filesystems." error message. Then I tried "smbumount xxx" and worked. I couldn't see the files in xxx directory anymore, and I could even delete xxx directory.
But now, when looking at /etc/mtab file, I still can see the share listed as mounted! The line is: "//server/share /home/someuser/xxx cifs rw,mand,nosuid,nodev,user=someuser 0 0". Also, df command lists the share as mounted too. But, as I said, the share is not really mounted, and I can even delete xxx directory without any problem... /proc/mounts doesn't list the share as mounted either... And, obviously, fstab file has no entry for this mount.
Moreover, /etc/mtab file group changed from root to someuser!I don't really understand what happened, but I assume some bug with smbumount or smbmount. Anyway, what does really worry me is how to fix this mess... I already changed mtab file group back to root, but I don't know how to return the mtab file to a consistent state.My main question is: can I simply edit mtab file and remove the inconsistent line?? I know one should never have to manually edit this file, but can it really be done in a situation like the one I'm describing? What could happen if I do it?
Also, I'm worried about if I don't do something to fix the inconsistency I could have problems when trying to shutdown the computer, or in any other situation. So, should it be safe to leave things unfixed and let everything return to normality when the system is rebooted?
after some hard times, I finally managed to have my Orange Huawei E1752 HSPA modem running (a USB modem enabling you to surf the net) the only thing is that I have 2 "problems" with this: I cannot find a clean way to eject the stick when I am done using it. when I want to run it, I plug it, it asks me for the PIN code, which I give, then I do the following manipulations: I first run:
usb_modeswitch it "creates" 2 files /dev/ttyUSB0 and /dev/ttyUSB1, what I check with: ls -l /dev/ttyUSB*
then I use the usbserial module to handle the stick: modprobe usbserial vendor=0x12d1 product=0x141b then I check that the stick is working with: dmesg|grep GSM that outputs something like: USB Serial support registered for GSM modem (1-port) option 1-5:1.0: GSM modem (1-port) converter detected usb 1-5: GSM modem (1-port) converter now attached to ttyUSB0 option 1-5:1.1: GSM modem (1-port) converter detected usb 1-5: GSM modem (1-port) converter now attached to ttyUSB1 option: v0.7.2:USB Driver for GSM modems finally I choose the "Orange Default" in the list of available connections in the network manager icon on the panel and I am finally connected The thing is, that when I am done, I just go to the same list and then click "disconnect" I go on "Computer" and left click on the SD 4Gb storage (the SD card also in the stick) safely remote drive or something like this. Then I thought that unmounting the device would work,
As a newbie (and especially regarding the use of those modems) I am running out of ideas to "eject" it cleanly, I also checked in the /dev/ file to see that the ttyUSB0 and ttyUSB1 files are present
A client is having a problem with a recently installed CentOS 5.5 x86_64 Virtual Machine (under VMware ESX 3.5). The problem is reported by the user as "Nautilus windows closing after a time when looking at data from the NFS server".
The folders the user is using are in a file system that is NFS mounted from a Solaris 10 file server cluster (VCS). These are mounted using autofs. These are direct mounts as in:
I have reproduced the problem by logging in and opening a number of folders using Nautilus under /opt/user-data. After a while, all Nautilus folder windows that are under /opt/user-data will close and when this occurs you can see from the command line that autofs has unmounted the area from the file server.
This should not occur because Nautilus having a window open should count as the folder being in use and should stop autofs from unmounting it. I don't know if this is an autofs or Nautilus bug but I suspect it's a Nautilus bug in it not keeping some kind of access open on the folder. System information:
The system was updated at install time. /boot/grub/grub.conf has "divider=10" appended to the kernel line.VMware Tools are installed. NTP is also configured and steps the time to UTC at boot if required (maybe this is overkill). autofs and Nautilus versions are: