CentOS 5 :: LVM2 Clustered Mirror Monitor And Repair?
Apr 8, 2009
I am using LVM2 Clustering to mirror logical volumes between 2 different storage arrays ( iscsi target hosts). Everything is up and running fine, but I would like to be able to monitor the status of the mirror(s), and "re-sync" them if necessary.
I have attempted the locate the information from Google. From what I have read, lvdisplay -v should tell me whether the mirrors are in sync, but it either is not, or I am simply miss reading it. I have come also across information stating lvchange -m [ 0 | # ] should be used to add or remove mirror copies. Is the is correct / only way to "resync" a mirror, in the event that I can determine the mirror is actually out sync?
For those who are familiar with Symantec's VERITAS volume manager, I am looking get the equivalent information related to plexes from vxprint, the equivalent function from /etc/vx/bin/vxreattach -br, and the progress information from vxtask list. I hope the information is currently "search challenged", or I am simply not fully comprehending what I have found.
I "inherited" a cluster of 2 systems with XenCluster, based on CentOS 5.5. About five weeks ago I ran the (long overdue) yum update, which crashed in the middle of updating on both nodes, so I had to reinstall all those packages by hand and had to re-enable their startup. So both systems are already in the not-ideal state.
One system rebooted itself during my absence and now won't boot anymore. It seems to boil down to clustered lvm not being activated in the initrd and therefore no VolGroup00 available, which contains the root. The other volume group has good reasons to be clustered but the root does not need to be. But I can't fix it from a rescue CD, every time the error comes up that the volume group is clustered.The exact error display at boot is at
what do I do to either fix the startup to ignore VolGroup00 being clustered OR change the status of VolGroup00 to non-clustered?I notice the other system has this too so my best guess is that I'll have the same problem in a bigger way when this system reboots too. Hoping for no power failures...Relevant versions of stuff:
I have two ext3 partitions within an encrypted lvm2 volume. when i start up my system it says that there are 0.3% non contiguous blocks.
This is my steup:
When i want to repair with repair system from dvd it tells me that the repair and check operation for encrypted LVM devices is not supported. so how can i fix my filesystem?
I've done a search thru the forums on a recent headache but couldn't find anything. now we all know that windows installations knackers the grub boot-up, and it becomes modus-operandi to repair the boot loader, with a live distro. however, in my current dual boot system (vista.sp2/fedora12), i've done a clean install of windows7 over the current windows ntfs partition. as standard, the boot menu would vanish. but when i put in the live-linux cd, gparted now shows the fedora partition (sda6) as LVM2. i'm somewhat of a neophyte with linux, but fedora12, suse11.2 and ubuntu don't seem to yet support this file system.
since i don't have critical data on the fedora partition yet, i usually wouldn't mind reformating this part', but the wireless chipset on my laptop cost a fortune on cups of coffee and beer to install, and in addition i used the fedora12 FEL spin and want to keep the installation intact. is there anyway to make the LVM2 indentifier back to ext3 or 4 and repair grub form there (without loosing the current partition kernal etc.)?
What I am trying to to is set up a cluster as follows:
- There are four nodes - Nodes three and four have large amounts of storage that will be used for storing user data. - Nodes one and two will be running a couple of network services.
I'd like have nodes one and two to have a clustered filesystem just between them to store data for the services run only by them.is it possible to share one filesystem between two nodes and another filesystem between two other nodes? The cluster seems to try to make the storage changes across all nodes. They will not be able to spend the extra money to get the same storage in all four nodes. Here is what I have done. I am using CentOS 5.5. After installing CentOS, I did this:
...which I think makes sense because the cluster was supposed to create the PV and the volume group, but did not. I don't even think it tries. The first time it tries to do anything to another node is when I create a logical volume.
I'm trying just a generic yum update to bring the machine up to date, I'm running into problems with two packages and their dependencies, kpartx and lvm2
I've removed any additional repositories, and performed a yum clean all
I think my problem comes down to device-mapper, it's asking for 1.02.32-1 to install lvm2 however if I run a yum list device-mapper it only shows 1.02.24-1.el5 as installed, and it is not marked for update. I've tried downloading the rpm directly, If i try to install the rpm with yum I'm told that there is nothing to do, if I try to install directly I run into problems with conflicts with the already installed package.
The kpartx seems dependent on device-mapper-multipath
I used parted to create a partition inside the logical volume, and then merrily used that partition, which appeared as /dev/pv-whatever/lv-whateverp1
Of course I created the FS as ext3.
So, after a reboot, I can't access anything in that logical volume with standard tools, as /dev/pv-whatever now only has the lv-whatever special file inside.
I can look inside the LV with parted fine, but parted can't copy ext3 filesystems.
Is there any way to get the data out of a partition created INSIDE a logical volume if that filesystem is ext3?
Trying to do a yum update to get everything to latest, towards the end it says this:
[Code]...
how to get around this? I tried yum clean all, then yum update, again but it did the same. I had other deps missing on other servers but yum clean all fixed them -- can't find anyone else who's had this specific issue either, nor an rpm called 'device-mapper-event' or the other things that are missing - am kinda stuck!
I have read that when using xfs with lvm2, prior to kernel 2.6.29 write barriers are by deffault disabled.As i want to migrate /home to xfs (and create a external partition to hold some data in xfs too) i was thinking in compile a custom kernel 2.6.30 or higher from [URL].I have read the wiki, all the warnings , and that for compile a kernel.org kernel you should use the Linux Kernel in a Nutshel guide/book. But i don't really know if would worth do that only to have write barrier support , more when in one month i will buy a ups to be sure i can shutdown the pc well if the power goes out.Should i enable write barriers on xfs with lvm2 ,although going to have a ups ?
i want install goldendict and source downloaded and i want compile it. for compile i need this packages :
Building under Linux: Make sure you have those dependency packages installed: libvorbis-dev, zlib1g-dev, libhunspell-dev, x11proto-record-dev, qt4-qmake, libqt4-dev, g++, libxtst-dev, libphonon-dev. They can be named slightly different in different distributions
I have a server, running Centos 5 and it is working fine. Now I want to have another server (server2), mirror of my current server (server 1) if any time my (sever 1) stopped responding, (server 2) should start working until I fix server 1.
I created a local mirror with a DVD Image, is there a way to get this mirror to 5.4 now?I already updated the updates/extras repo, but now my main-repo is outdated and my updates go wrong.my directorystructure looks like this:
/server/rpmrepo/centos/5] # ls centosplus extras os updates
I have create a LVM volume in mirror 1 method, with two hard drives, In that one disk is get failed. How can I manage the data availability and also what I need to do next?
I am trying to create a local mirror/repository for CentOS 5.6 for updates and network installs. I have read the page here on the site about how to do it over and over again, and I still can't figure it out. I have already created the directory, but right after that, I can't go any further. I know this is much easier than I probably think it is. Would someone mind telling me, (in beginners terms) on how to do this? I just want the 'os' package and nothing else.
I'm looking for a way to have two hosts talk to the same volume without the need for backend host communication ie. clustered file systems? Does anyone know of a clustered file system that will allow this? The basic idea I have in my head would be turn off caching and utilize stubs or pointers to let each host know when to reread new data? The hosts will not be writing to the same files or directories at the same time. They will also be targetting certain information in certain locations on the volume.
I'm trying to join two enclaves to pass data without networking. One host will drop a file and write a stub saying the file is there. The second host will then pick up the file analyze it and then drop a report so that the first host can then pick the report up.
I am Setting up 3 HP Proliant servers with 2 250GB Hard Drives in each. How Do I mirror the drives, and link all the servers together??? I am installing Centos 5.3. From the software install menu I installed Gnome, and the Server software, but can't seem to figure out how to get into the server side of it from gnome. Sorry for all the questions, but I have been out of the loop of all this for about 8 years and trying to re-learn everything.
I recently setup a software RAID1 with two hard drives for testing. I manually unplugged one hard drive to test redundancy and it was fine. I then did the same to the second and plugged back in the first and it was fine. But when I plugged both back in it is only loading one of the mirrors instead of both. I am guessing when the mirror (failed), at least as far as the system knew it removed it from the load sequence to prevent data corruption... how do I go about re-enabling it?.
cat /proc/mdstat Personalities : [raid1] md0 : active raid1 sdb1[1]
I fouled up the file system when removing a drive. How do I fix it, or do I need to re-install?
The system boots to the point in the GUI where it checks the file system. It then suggests that I run fsck without the - a and -p options, and the drops to a sheel. I enter the root pwd and then it says: "(Repair fileystem ) 1#" What do I do from there?
i know i asked earlier but i got my usb to be mounted on a different fedora distro. how do i now find my documents or repair it? also how do i add in a repair line if possible?
I was puzzling over the nss dependence problem that people are having in upgrading 5.2 to 5.3. The issue is clear (that the mirror people are using for [updates] is not pointing to the latest set when an uptodate mirror may be being used for your [base]). My question is though how the mirror list decides whether a mirror is fresh. According to [URL] then this mirror [URL] is "green" which presumably means that the system thinks it is up to date (last probe was 1 hour ago). However if you look at the files on the mirror now (12:54 BST, Apr 1 2009) you see that the date of the 5/ branch is 24-Jun-2008. Thus this host is not ready to give you updates to 5.3. Is this a bug?
I've just set up the local mirror for my 96 CentOS 5 workstations. The mirroring script is taken from public-mirror howto and looks like this:
#!/bin/sh rsync="/usr/bin/rsync -avHzL --delete --delay-updates" # replaced -q with -v for debuging purposes, removed --bwlimit, added -L to follow symlinks mirror=centos.politechnika.lublin.pl::CentOS # tried several different same result
[Code]....
I tested mirror consistency by putting the mirror addres I'm syncing with, directly into repo file. Than yumex works fine. I tried several different mirrors with the same result.
After reading a lot of docs, I'm still having problems using wget to download a Centos repo from a mirror. Here's my best attempt so far: $cd /repos/centos/5.4 $wget -r -nH --cut-dirs=3 -np [URL] Of course I get all the unwanted index files etc, but I seem to get a lot of other downloads from the mirror, not just their 5.4 directory. It's like it's following other links on the web pages. Maybe I should be using "ftp://" instead of "http://" considering it's an ftp site, but I seem to have connection problems that way.
I have created two logic disks /dev/VolGroup00/LogVol00 and /dev/VolGroup02/LogVol02 After reboot has disappeared /dev/VolGroup02/LogVol02 How to restore logic volume?
since yesterday, this domain has been painful slow and the only fix is removing it from /var/cache/yum/updates/mirrorlist.txt the last straw was about an hour ago, when 10kb/s was the max bandwidth and I had to download perl and glibc * glibc-common it would have taken over an hour. once removed from updates/mirrorlist the other mirrors picked up the slack and I was back in business,
So everytime i try and use the "yum" command. for some reason it doesn't go past this point: "loading mirror speeds from cached host files". i cleaned up the cache and rebuilt the OS again, and im still getting this problem. would this be a problem with my internet connection? I'm using CentOS 5.5.
We have two CentOS 5 servers in production (web and database). We are setting up a single staging server that will mirror the configurations of these servers as closely as possible. What is the easiest way to ensure the exact same software and configs as the production servers are setup on the new staging server. Our contracted data center provider has already informed us that they do not perform images and NO we do not have physical access to the machines. It is undetermined whether we will be virtualizing the staging server into two virtual servers yet, so for the purposes of this post lets assume we are not. I'm seeking a faster/more precise method than doing this by eye and hand.
