Server :: Network Disk Mirroring Of >16TB XFS Volumes?
Jul 29, 2011
I'm setting up two media servers - one will be primary and the second will be the failover box if the first one dies. My original plan was to use Heartbeat and DRBD to replicate the first to the second, but it turns out DRBD has a limitation of 16TB per volume, and my RAID 6 is 30TB in size. I'm trying to figure out some options.
One is to split the partition into two, so it's under the 16TB limit of DRBD, share both partitions, and then use LVM to join them as a single volume at the OS level. I'm not sure what kind of performance hit that will make on read/write access since now it will need to go through LVM *and* DRBD.
Two is to use something like GlusterFS or XtreemFS. However, I am wary of both of those as the first is a FUSE-based system and isn't really tuned for performance, and XtreemFS is still kind of beta-ish.
Three is good old rsync, but that is a bit too asynchronous in my opinion. I'd like to be able to fail over virtually instantly and not lose any content.This is a production system and performance of the disk systems are paramount, followed by reliability.
pls suggest me how to mirroring two hard disk drive in rhel9 server.i can do raid on those hard drive but user requer mirroring.so pls help how could i do this, mirror two hard drive.
I am using Ubuntu 10.10. I have a system set up with 1tb HD. I also have another 1tb HD which I'd like to use to mirror the other drive. So if the primary HD fails I can boot and operate from the mirrored drive. I've read that this is possible by using Raid. however I am confused if it is possible to set-up with a HD which is already set-up Ubuntu system. Also what what I can make out the mother board does not have a raid option.
I am looking for such a tool, very much like the one that is on Windows 7, where you can basically with one glance see how much of a volume is occupied (graphical bar) and that for all mounted volumes. I have been looking for this, but so far I have not found it.Also, important: that it is auto updated. So that it is not like a report that was generated and then does not change anymore, but a live thing
I have a mail server on Fedora 8. I would like to have a mirror to this server so that it takes over in case my server goes down.I dont know how to go about it.
I have a huge RAID6 array 21TB+ already partitioned in GPT. This is to be used as the storage location for my company's backup server, and I want to access it as one logical volume. Is this possible with Centos5? I just discovered the product specifications for Centos5, and saw that the maximum file system size is 16TB, but LVM2 should support volumes up to 1EB. Is there any way I can make this work in Centos, or am I going to have to run a different distro?
I have two web servers. One is active and one is in reserve. I keep the user data (web pages) in sync by running rsync every 10 minutes or so. This copies any changes from the active machine to the reserve machine. But, it's slow, only gets changes every 10 minutes, bogs down the disk, does strange things to files that are changing during the rsync process etc...
I want something that will automatically copy any changes from the active server to the reserve server as they are made. IE I hit 'save' on the active server, it copies the file to the reserve server. Simple!
I've been looking around and I see GFS which is really vastly more complicated than I need. I'm happy with read-only access on the reserve host, so I don't need distributed lock management.
I could theoretically implement this by setting inotify watchers on every file and running an SCP or rsync command when a file gets saved. So, it can't be that hard.
I do not need a true networked file system, as in something I mount over the network. I just want something to keep my disks in sync.
Blue sky thinking at the moment:I have a number of file servers, each at different sites. I would like to be able to make these appear as one, so that files on any server can be accessed from any site, and the user doesn't even see there are multiple servers.Obviously, the internet is slow, especially the upload speeds. So when a file is written the write ought to go to the server on the client's LAN - even if it was previously on another of the servers.However, for robustness, some sort of background mirroring is also wanted. If all the servers were left on and connected, they eventually end up all in sync. But this mirroring needs to be mindful of bandwidth usage; if someone writes a big file to their local server, copying that to the other servers can't interfere with normal internet usage.I think UnionFS or similar might be able to handle the unioning side, but not the mirroring stuff.
I want to setup a Linux File Server for a small windows network (around 50 users). I do know that I am gona need Smb service/pkg for that. I haven't used Samba for a while now and as per the best of my knowledge, entire communication (including usernames and passwords) between a samba server & windows client machines will be plain text. Is there any way to secure all this communication??
Secondly, if i remember correctly, MS windows wont let me mount more than one samba shares as network disk when all my shares can be accessed by different smb users with different passwords?? is there a solution to this problem? OR may be if there is any other package available for this purpose so that i wont have to use samba?
I have been asked to use space allocated to /dev/sdb2 to grow /dev/sdb1 both filesystems contain data. When I run lvdisplay or vgdisplay I get "no volume groups found". and pvdisplay returns nothing. Do I need to add the fs to a group?
I'm using munin to monitor our servers. Recently I created additional partitions on one of them and I noticed that munin has not picked up the changes.
I restarted munin-node and running the plugin manually as user munin on the machine
Code: $ munin-run df I get all my partitions Also when I telnet 10.1.1.25 4949 and run
Code: fetch df I'm getting all of them. The graph however displays only the ones that were there at the time of munin install.
How can I get munin to add the new partitions to the graph ?
I'm trying to make a fresh install of Ubuntu 10.10 Server 64bit but is stuck with the Partitioning Disk part and Ubuntu 10.04 LTS provides the same error. The installation wizard gives me a menu:
!! Partitioning disks This menu allows you to configure iSCSI volumes iSCSI configuration actions
We have a Centos 5.6 server mounting two iSCSI volumes from an HP P2000 storage array. Multipathd is also running, and this has been working well for the few months we have been using it. The two volumes presented to the server were created in LVM, and worked without problem.We had a requirement to reboot the server, and now the iSCSI volumes will no longer mount. From what I can tell, the iSCSI connection is working ok, as I can see the correct sessions, and if I run 'fdisk -l' I can see the iSCSI block devices, but the O/S isn't seeing the filesystems. Any LVM command does not show the volumes at all, and 'vgchange -a y' only lists the local boot volume, not the iSCSI volumes. My concern is that, the output of 'fdisk -l' says 'Disk /dev/xxx doesn't contain a valid partition table' for all the iSCSI devices. Research shows that performing the vgchange -a y command should automatically mount any VG's that aren't showing, but it doesn't work.
There's a lot of data on these iSCSI volumes, and I'm no LVM expert. I've read that some have had problems where LVM starts before iSCSI and things get a bit messed up, but I don't know if this is the case here (I can't tell), but if there's a way of switching this round that might help, I'm prepared to give it a go.There was absolutely no indication there were any problems with these volumes, so corruption is highly unlikely.
what the maximum number of logical volumes is for a volume group in LVM ? Is there any known performance hit for creating a large number of small logical volumes vs a small number of large volumes ?
I have installed fedora-ds-1.0.4-1.FC6.x86_64.opt.rpm on RHEL 5 Enterprise, which I use for Oracle BIEE. Now I need to install the FDS in another RHEL 5 and mirror only the FDS.
configure the mirroring part for the Fedora Directory Server.
I am having problem with the root disk / fill up. So now that i have the disk information provided from SAN team Id 0:3, how would i do to add this device to / to increase space on Redhat Enterprise server 4?
I have several SLES 11sp1 servers and using SMT (Subscription Management Tool) to provide local mirroring and prevent all the unnecessary web traffic to update all these machines.
I found this article which is working out quite nicely to mirror the openSUSE 11.3 Updates which prevents all my openSUSE 11.3 machines from all going out getting the same update.
Code:
[URL]
I am looking to try and do the same with openSUSE 11.4 and with openSUSE 11.4 using the Tumbleweed repos. Has anyone done this or can someone provide some insight on what I need to do. I am thinking I can run this command
I should use* to provide me with openSUSE 11.4 Updates. For the Tumbleweed repo would I do the same and would that work?
Code:
smt-setup-custom-repos --name OpenSUSE_Tumbleweed --description "OpenSUSE Tumbleweed" --exturl http://download.opensuse.org/repositories/openSUSE:/Tumbleweed/standard/ --productid *don't know what productid I should use*
Again I don't know what productid I should use or if it really even matters, could I use one that is not being used? I am trying to limit how many machines go out to the Internet to pull down updates. I would really like to have one system (SMT) pulll down any new files and all my clients get their updates from the SMT server.
I've got a repository (3rd party, but not a launchpad one) in my APT sources list that takes a long time to respond on apt-get update.What I was wondering was if I could keep a local mirror of just that repository, and have apt-get update refer to it instead of the normal one, then leave a cron job to check the normal one for any updates.
I've set up a test machine with Ubuntu 10.04. It has two drives making one RAID 1 drive using mdadm. This RAID drive is where /home is mounted. I like to break things just to see what happens and to know what to do to resolve it before it happens for real. So I physically removed one of the drives that made up the RAID 1 (while it was off).
I then rebooted the machine. I thought since it was mirrored then /home would mount correctly using the other mirror. What actually happened was on the Ubuntu splash screen it said there was a problem mounting /home. I skipped the mounting, logged in and looked in /proc/mdstat. It reported one drive that was inactive it did not report the missing drive.
Currently I backing up the MBR, the C and the other partitions to an external USB HDD and from there I restore them if needed. I use the SystemRescueCd and commands like dd if=/dev/sd* of=/mnt/PC_name/backupmbr.1 count=1 bs=512 and ntfsclone --save-image --output /mnt/PC_name/PC_name_c.img /dev/sd*1 etc. I want to clone the HDD the way, however, that I omit the external USB drive. I want to connect the new HDD to the PC and do the cloning directly from one disc to the other.
My questions are:
- Can you provide me with the exact command? - Is that a difference if the disc is SATA or IDE? - Can I copy the disc even if the old disc don't wan to boot?
I've always been under the impression that when using the console only a notebook's display is used. Apparently this is not the case as I just booted off of a SystemRescueCD and the display on the notebook was mirrored on my external HDMI connected monitor. I have a feeling this is set somewhere in the kernel, but where?
