Debian Hardware :: Speed Up RAID5 Recovery From WD EARS HDD
May 15, 2011
1 - I got the array up and running on three Seagate 1.5TB HDDs, filled it up in 6 months (mostly by putting all of my stuff in one place)
2 - Unknowingly, I bought a WD15EARS HDD, partitioned it as I would any normal drive and asked mdadm to grow my array over to the new partition
3 - Worried by abysmal read/write speed, I investigated, found out how to deal with the new HDD ( here or here )
4 - Being a newbie at this, I opted to do a trial run before wiping the disk and unmount/mount the partitions, LVM layers, SW RAID array. This is there I messed up.
I am hoping someone can help me I setup a brand new ubuntu 11.04 64bit system with:-AMD Phoenom 840 (3.7GHz overclocked) 8GB DDR3 RAM.WD20EARS x 4madmin (RAID5) SetupFor some reason I cannot get the write speed above 15mb/s & I have tried modifying the cache_Stripe_size from 256 to 8192.My read speeds are great burst at 120mb/s & stable at about 60mb/s.The only other possiblility is it could be my HDs which I read causes slow writes not sure if thats true though. point me in the right direction so I can fix this issue as I don't want to move to the dark side (Windows) .
I had a sata drive go down on my software raid array a couple days ago..After replacing the defective drive, I rebuilt it using fdisk to set partitions and mdadm to add the drive back into the raid array.After getting md2 to mount to /mnt I copied the /mnt/etc/fstab and /mnt/etc/mdadm.conf to the /etc directory and rebooted..When I boot, my /dev/md2 partition is not mounting to / It is not loading all of the services at boot time it should.. the size of /dev/md2 on / is even wrong.. it should be about 1.3TBwhen I manually mount /dev/md2 to /mnt I can access all of my data, through /mnt/var/www/vhosts and so on.. I am at a loss, why is the machine booting to this craziness?All of my data, that I know of can be accessed when md2 is mounted to /mnt[root@ESS000272 etc]# cat mdadm.conf
# mdadm.conf written out by anaconda DEVICE partitions MAILADDR root
Is there a thread on the new Advanced Format standard? I'm reading that it's pointed at W7/Mac and I'm thinking Linux newbies (like me) should know in advance, before buying hard drives, what the state of OpenSuSE/kernel/drivers is. I never used to pay attention to hard drive numbers.
My System Intel Core Duo E5300 Mobo - Gigabyte G31M-ES2L 1GB DDR2 800MHz RAM 4 x WD20EARS HDD I have been trying to install Fedora or Ubuntu for over a week. I thought it would take an hour and i would be away. I have been trying to install using the mdadm Software RAID feature. Everytime it takes about a day to format the drives and then i get an installation error. The drives state they are ready to use as is on any operating system other then WinXP, but this does not appear to be the case.
I am very new to Fedora... I have been doing some reading.[URL].. That information has been promising. I have been able to get into Fdisk off the live CD but i can't figure out how or if it is possible to do what i want it to.
Has anyone had any luck getting these drives to function correctly in a software RAID? I have had good luck with WD drives in the past and just assumed these drives would do what i wanted to but alas i have been proven wrong.
The partitions i wanted was... - A 2GB swap parition - A 10 GB RAID 1 partition for Fedora - The remaining space as a Raid 5 for files.
Am i just banging my head against a wall here, or is this possible.
I have no drive failures but just need to recreate a raid5 set as the next free MD disk number. Originally I built a temp OS of debian on a single drive and had 4x2TB drives in a raid5 software array (MD0) this worked fine and allowed me to move all data to it, and remove our old fileserver. I have now pulled out the 4 x 2TB Raid 5 drives and created a new OS on two new 80GB drives, partioned as follows,
MD0 is now 250mb Raid1 as /boot MD1 is 4GB Raid1 Swap MD2 is 76GB Raid1 as /
If I turn off and push back in the 4x2TB drives I cannot see a MD3. I presume I would need to create a MD3 from these 4 drives but I dont want to mess things up as its live data. So im here asking for help, or a bit of hand holding to get it done right.
PS - Its a Debian Lenny 5.0.3 Raid1 fresh install replacing a Debian Lenny 5.0.3 on a single disk.
i was adding another disk to my raid 5, all was going well it started the reshape, got past the critical zone, worked for 20mins, but now it seems to have crashed.When i cat /proc/mdstat, or mdadm -D /dev/md0, those programes hang and dont print anything or return.from my kern.log i can see that there was an error on a disk, the raid array removed it, was going to continue the reshape but finished immediately. Anyone know what i should do?
Mar 6 16:20:26 Aries kernel: [1931119.599107] md: reshape of RAID array md0 Mar 6 16:20:26 Aries kernel: [1931119.599107] md: minimum _guaranteed_ speed: 1000 KB/sec/disk. Mar 6 16:20:26 Aries kernel: [1931119.599107] md: using maximum available idle IO bandwidth (but not more
Software RAID 5 on 4 2TB caviar green disks. Then 12 partitions or about 500GB, and finally LVM to use the partitions as pv, and create some vg for my needs.
My disks are /dev/sda to /dev/sdd I boot on /dev/sde which is an sd card
I've got a, as it seems to me, strange problem.I've inadvertently deleted my user from the group admin so I'm in the same situation of a lot of other users (read a lot of messages about it).My problem is that when restarted in recovery mode there is no way I can choose the 'drop to the root shell' or similar in the menu.The menu appears for a second and then I've got an empty screen. If I press a key I've been requested for a username and password that of course is not what I need.
I received the following error when I got home from work today. If this was a windows environment, my first inclination would be to boot off my dvd and then run a chkdsk on the drive to flag any bad sectors that might exist. But there's a complication for me.
Code: Select allThis message was generated by the smartd daemon running on:   host name: LinuxDesktop   DNS domain: [Empty]
The following warning/error was logged by the smartd daemon: Device: /dev/sdc [SAT], 1 Currently unreadable (pending) sectors Device info: WDC WD5000AAKS-65V0A0, S/N:WD-WCAWF2422464, WWN:5-0014ee-157c5db9a, FW:05.01D05, 500 GB For details see host's SYSLOG.
You can also use the smartctl utility for further investigation.The original message about this issue was sent at Sun Feb 14 13:43:17 2016 MST.Another message will be sent in 24 hours if the problem persists.
From gnome-disks Code: Select allDisk is OK, 418 bad sectors (28° C / 82° F)
I did a bit of reading and it seems that most people suggest using badblocks to first get a list of badblocks from the drive and save it to a file. Then use e2fsck to then mark the blocks listed in the badblocks file as bad on the hard drive. My problem here is that this drive is part of a RAID5 array that hosts my OS. I wanted to confirm if this was still the correct process.I boot to my Live Debian disk, stop the raid array if it's active. Then run badblocks + e2fsck commands on the drive in question and then reboot.
After a large update of software I can nolonger open Squeeze. I get past the grub screen however before the gdm starts up the screen goes blank and stays that way. I have only one kernel showing on the grub screen.
I tried in Recovery Mode: Ctrl+Alt +F keys with effect and startx returns a blank screen, this is why I though to try loading another kernel to see if that helps. I'm using a Acer Laptop.
My question is: Is it possible while in Recovery Mode to load/install another kernel os that when I get back to the grub screen I have another kernel to try to boot into?
I have just upgraded my lenny box to squeeze. I did it by clean-installing squeeze. The installation was successful, but I just noticed that I had forgotten to backup some important files I had on this machine before the installation...
I have inherited a Compaq Armada 1592DMT that has Debian GNU/Linux (tiny) on its hard drive. After bootup there is a dialog box asking for login & password. Two things I have no way of answering. So how do I recover them? I do not have the faintest idea how Debian Linux works, nor how to Bash Grubs, I'm just an old Windows button pusher.
I'm a beginner + in debian (ie I know how to launch basic command under a shell), and I very new to configuration problem.I have a running perfectly file server for probably more than 2 years that I never touch. That's the debian problem, when it's unning, it's forever, no reboot at all.I recently bought 2 new sata drive 1TB5, that I add to my system by adding to my old system a SATA PCI board without any problem.But, one of these 3 month old drive is already dead.So I try to replace it and that's where problem started.....First I did a#mdadm --manage /dev/md1 --remove /dev/sdb1, of course without sucess because sdb is dead
#mdadm --detail /dev/md1 /dev/md1: Version : 00.90.03
had a bad experience when Fedora 15 overwrote the ext4 partition of a data disk to MVL during the installation process.I cloned the HD and now I am working on it. However, my first attempt resulted in 900.00 number-renamed files into the lost+found folder. And that's not what I want: with this number of files I need to recover the directory structure and the files real names.I know this is a hard issue for being discussed in the forum and that I shall look for some expert help, but, I wished to useis bad moment as an opportunity of learning.
Ubuntu 64bit. The sound system works and plays noises correctly when I test the speakers in sound preferances. The internet BBCi player(Radio) plays sound correctly. Banshee & Rhythmbox try to play music files at double, or more, speed with no sound output. Spotify Linux version also tries to playback at double speed with no sound output. Media Player attempts to play music files at high speed. Media player plays the Video and audio tracks at high speed. VLC Will play the video at normal speed but with no audio.
After recovery from a disk crash, everything seems fine except that I have lost plasma. I have a backup. Is there a way to reload plasma from the backup? I can recover kde by reloading /home/.kde from the backup. But there is no .plasma file.
I've got an 8-disk raid-5 setup, and one of the disks failed. I shut the system down, replaced it, and powered the box back on again. Then, I made a catastrophic mistake; I 'failed' and removed the wrong disk (should have been sdj1, and I typed sdk1 by accident). I tried to re-add sdk1 back to the raid array, but it got listed as 'spare'. My raid array is off-line, since I now have 2 disks unavailable.
I know that the data still exists on sdk1, is there any way I can get the raid array to recognise the fact that it's a valid part of the array, and not a spare disk? At least if I can do that, I'll have a degraded but accessible array, and then I can rebuild the array on the properly replaced disk.
i both have the same problem, i'm trying to burn my images at 4 or 8 speed, but ubuntu 10.04 says that the hardware does not support that kind of speed and switch up to 16 speed and more. i know it can burn at low speeds, at least in windows, it is a bit strange that fast burning is okee, and slow not, what can i do to prevent this? i don't wanna burn to much errors on my discs
My previously running debian system recently started failing with systemd failure messages as some services are being started. It never completes. I can log into recovery mode.
I added systemd.log_target=kmsg systemd.log_level=debug into my grub parameters.
After I do that all that I see scrolling down the page on boot are:
systemd-journald[36] /dev/kmsg buffer overrun, some messages lost and the occasional systemd[1] Looping too fast Throttling execution a little
Is there a way that I can boot without systemd using sysvinit?
I have added to Debian Wheezy testing repositories. After a restart there is no desktop anymore just the wallpaper picture is visible but no desktop pop-ups.
When I want to install the desktop through recovery mode, I get an error message that it "can't fetch...repos". How I can repair /etc/apt/sources.list through recovery mode?
I would like to build an oem style install partions that is bootable with menu to choose if I want to run install or boot already installed system. I would like to include current source packages on the same dive so if I don't have internet access at time of install, can can still install what I need.I know with Windows Vista and Windows 7, you can get this but how can I do this with Debian?
I am experiencing the slow transfer speed from my Lenny to Windows 7. Both of my NIC's are oboard Nvidia nForce 1GBPs. For example 200MB transfer from Win to Lenny (default samba shares) takes about a minute, while the other way around takes an hour!!!. I get similar transfer behavior through sFTP.
Can somebody clarify if the problem is in NIC drivers, MTU (I can't set it up higher than 1500), or smb.conf settings
I have a 500gB USB drive connected to my laptop for backups and filestorage. But I can't get it to play nice with Midnight Commander. My transfer speeds max ut at 2MB/s wich is painfully slow when moving large files such as movies. Worker FM transfers the same files to the same drive much MUCH faster (not sure by how much, though). This leads me to the conclusion that the problem lies with MC.
I wanted to play with Kali Linux, so since my T430 has two USB 3.0 ports I bought a USB 3 thumb drive and put the image on it.It boots fine and all, but I recently read that the T430 downgrades the USB port from 3 to 2 when you boot from a device connected to it.I followed instructions here: URL...
Code:
Select allroot@kali:~# dd if=/dev/zero of=./largefile bs=8k count=10000 10000+0 records in 10000+0 records out 81920000 bytes (82 MB) copied, 0.0670296 s, 1.2 GB/s root@kali:~# sh -c "sync & echo 3 > /proc/sys/vm/drop_caches"
Are there any open source or third party applictions can could do a bare metal recovery on Debian or any linux machines? We are looking for a solution that won't need a shutdown or reboot.
Is it possible to control the speed of the CPU fan in my desktop computer? The CPU runs cool at 30C yet the fan sits around 2700rpm and the BIOS doesn't allow for speed control.
