Tag Info

Hot answers tagged

3

You have four primary partitions and want to add a fifth... and you can't just redeclare them extended/logical because those need an extra sector for each partition. Also GPT has a backup at the end of the disk so if you ever lost a partition table to MSDOS and had to resort to TestDisk, with GPT you might be able to do without. grub on BIOS system with ...


2

It is by design that the kernel tells userspace utilities that a file operation completed successfully while the caches are buffered. Try editing /etc/hdparm.conf to disable write caching with the option write_cache = off. However this will reduce performance, and it is usually preferred to use sync if you want to ensure write buffers are flushed to disk.


2

"ICRC error" means that data was corrupted during the transfer; given the age of your drive, the most likely cause is a bad cable. All told the possible causes seem to be: a bad cable (try re-seating your cable, and changing it if that doesn't fix things); a bad chipset (unlikely if your motherboard isn't very old); a bad drive. If you get to the last ...


2

I'm not sure that you can get all of this information from the ext4 driver, because it does not responsible for the disk sectors and other disk geometry, but block device layer sub-sytem in the Linux kernel. I don't see one way to get all information in which you are interesting, because as I see you are in teresting as in low-level stuff and also in high ...


1

You need kernel level tracing to achieve that. There is a large number of tools available to do it with various levels of features, usability and stability including sysdig, ftrace, perf, dtrace4linux, ktap, systemtap and others. You might start with tpoint which, being based on ftrace, shouldn't need anything to be installed (outside the script itself), ...


1

Your question is unclear. If you want raw disk I/O profiling (then file system is irrelevant, and you want also to measure swap disk IO) you might have to configure or patch the kernel (I don't know how). Perhaps running Linux thru an hypervisor (like xen) might help. Look also into oprofile On the other hand, if you are interested by file system activity, ...


1

GETGEO returns bios drive geometry, which is obsolete. IDENTITY returns the raw ATA device identification sector. You shouldn't use either one. Instead, simply read from the files /sys/block/sda/size and /sys/block/sda/queue/hw_sector_size. The former gives the size in "sectors" as if the sector size were 512 bytes, even if it isn't, and the latter gives the ...


1

It depends upon the hardware connection between the disk drive and the computer. But if it is an USB 2, indeed it would be slow (whatever OS you use). The alternative could be to add a new hardware SATA drive, or to repartition your existing internal drive (to have some space for your Linux system). Perhaps you could repartition (& shrink the existing ...


1

NTFS is a proprietary file system (MS). Everything ntfs-3g is able to do was achieved by reverse engineering. Considering the above, I would not expect a proprietary file system to be as fast as an open file system (under linux). Now, if NTFS should be only 20% slower than ext4 (instead of 50%) in your case, i do not know, but i'm sure you can have a clean ...


1

As I commented (and for obvious performance reasons) the kernel is using a page cache. So this is a feature, not a problem. See http://linuxatemyram.com/ for more. You could (but I don't recommend doing that) using some mount options (to disable, or lower the use of, the page cache), and you need to umount any device (e.g. an USB key) before unpluggging or ...


1

Mount /dev/md2 first as something like /srv/DataNew, run a 1st round of copy as root (actually I'd suggest rsync, IMHO it's better for this kind of job): rsync -a --delete /srv/Data/ /srv/DataNew Optionally you can re-run the cmd - the 2nd execution should be faster (rsync is capable of skipping files already copied and up2date) and will give you a rough ...


1

In a word, yes. More specifically, you need to mount the new array somewhere else ( can't mount two things in the same place ), copy the files, unmount both, then set your system to mount the new raid in that location ( edit /etc/fstab ).


1

I successfully achieved converting MBR to GPT, but used two extra (new) disks for safety reasons. Note that I'm using Debian in combination with the GRUB bootloader. With my setup, which simply has a Linux partition and a swap partition, the procedure is roughly as following. First, make a full backup: use the first extra disk to make a full backup of ...


1

In my Alpine Linux Partition scripts I use: check_scheme() { fdisk -l $1 |grep "Disklabel type:" |awk '{ print $3 }' }



Only top voted, non community-wiki answers of a minimum length are eligible