Sponsored Content
Top Forums UNIX for Advanced & Expert Users VxVM breaking mirror for migrating storage Post 302895551 by keelba on Tuesday 1st of April 2014 09:24:00 AM
Old 04-01-2014
VxVM breaking mirror for migrating storage

We are in the process of migrating storage from one array to another. I know the commands I would use in LVM but I am not very familiar with VxVM.

The goal is to attach the new LUNs from the new array and add them to the disk group. Then mirror the data online to these new LUNs using VxVM mirroring. Once the mirroring has completed we would break the mirror from the old LUNs and be running on the new storage.

The catch here is that the customer is not 100% sure they want to be on the new array and wants the ability to failback quickly if there are problems, such as performance. There are several terabytes of data so the original sync could be quite time consuming. If, for example, it takes 12 hours and the customer complains about the performance, we want to get back on the old disks as soon as possible, not after waiting another 12 hours.

In LVM, I could do this simply with the lvsplit and lvmerge commands. The lvsplit breaks the mirror into a new lvol and keeps a bitmap of all the changes so that resyncing the mirror happens much more quickly, very similarly to a snapshot. If, after a couple of days, the customer is satisfied, we would completely remove the old LUNs and be done with the project.

Can anyone help me with the commands to do the same in VxVM? I found conflicting answers on Google. One says an exact equivalent to lvsplit/merge is the command vxassist snapshot and vxassist snapback. Antother sites suggests using vxplex detach and vxplex attach. So if anyone has suggestions on the best way to accomplish this, I'd really appreciate it.

One thing I need an answer for, is where the bitmap is stored and how to make sure we do not fill it up and lose it. This is an active database so the delta might be quite large for all we know after a few days.
 

10 More Discussions You Might Find Interesting

1. HP-UX

Breaking Mirror

Can some one point this UNIX newbie to a web site or directions on the steps needed to break a mirror in HP-UNIX to change a bad hard drive. (4 Replies)
Discussion started by: egress1
4 Replies

2. Solaris

Breaking mirror for patches, sun/os 5.9

Hi, Going to be applying patches to Sun prod (dev 1st) boxes soon, I'm going to break the mirror's on 3 file systems on the root disk prior to that (/, /var, /swap),then run 9_Recommended. Then, run it for a day or 2 before re-syncing the mirrors. But, what if I want to backout of this... (1 Reply)
Discussion started by: BG_JrAdmin
1 Replies

3. Solaris

Solaris 10 - breaking of mirror and change new hard disk

Hi, am a newbie at solaris. Need advice and help on this. 1) How do I break the mirror between 2 hard disks. (wish to keep 1 good hard disk as backup) 2) After remove 1 hard disk and put in new hard disk, how do I initialise or fomat the new hard disk? 3) How do I put back the backup... (3 Replies)
Discussion started by: chongkls77
3 Replies

4. Red Hat

Migrating from RHEL LVM to VxVM

Hiya, Is there a known method from converting LVM data to VxVM data.. ie we have a load of local and SAN (HPDM) mounts on RHEL 4u5/5u1 servers which are now required to be clustered with SFS (VxFS,VxVM,CFS,VCS etc) I've come across the vxvmconvert, but thought that was HP-UX only? Many... (0 Replies)
Discussion started by: itsupplies
0 Replies

5. Filesystems, Disks and Memory

Migrating from RHEL LVM to VxVM

Hiya, Is there a known method from converting LVM data to VxVM data.. ie we have a load of local and SAN (HPDM) mounts on RHEL 4u5/5u1 servers which are now required to be clustered with SFS (VxFS,VxVM,CFS,VCS etc) I've come across the vxvmconvert, but thought that was HP-UX only? Many... (3 Replies)
Discussion started by: itsupplies
3 Replies

6. AIX

IBM SAN Storage DS4700 - Copy Mirror Flashcopy

Hello, Does anyone know how to copy SAN Storage logical disks from IBM TotalStorage Software. I have a SAN Logical Disk of 200GB mounted on my AIX LPAR_1 via fibre channel fcs0 I would like to make an exact copy of the SAN Logical Disk from IBM totalStorage and mount it on AIX LPAR_2 ... (4 Replies)
Discussion started by: filosophizer
4 Replies

7. Red Hat

breaking a hard drive mirror

Hey guys i have a machine running redhat enterprise 3.8 (i believe) and it has two hard drives that are mirrored. what would i have to do to get them to act as two seperate hard drives so i have 120gb capacity total as opposed to just 60? also, what would i do if i just wanted to take one of... (4 Replies)
Discussion started by: pacman529
4 Replies

8. Solaris

Unable to boot from mirror disk on x86 server configured under VxVM

Hi, Can you help me on booting x86 server configured under VxVM. Server boots fine normally from both the disks but if I try to boot server from mirror disk without starting veritas, then it does not boot. vxplex -g rootdg dis var-02 vxplex -g rootdg dis swapvol-02 vxplex -g rootdg dis... (2 Replies)
Discussion started by: milindphanse604
2 Replies

9. Red Hat

Migrating an Oracle database storage - REDHAT4

e have a RHEL4 server with an Oracle 10 ( RAC ) are migrating storage bank records. The copy is made by the storage they make a copy of the level of disk all blocks are copied and the file permissions. Problems we had Initially the Cluster service was not rising due to absence of OCR disks... (0 Replies)
Discussion started by: fausto.ncc1701
0 Replies

10. AIX

Migrating LUNs to new storage array

All, I have a AIX server with multiple volumes and I need to move them from a legacy storage SAN to a new SAN. We are concerned because to use mirrorvg as we would have to run on a single HBA (one on the old SAN and one on the new). Our other option is to use an appliance to do a block... (3 Replies)
Discussion started by: bkchicago
3 Replies
lvsplit(1M)															       lvsplit(1M)

NAME
lvsplit - split mirrored LVM logical volume into two logical volumes SYNOPSIS
autobackup] suffix] PhysicalVolumeGroup] lv_path ... Remarks If the logical volume input arguments belong to a combination of volume groups version 1.0 and 2.0 or higher, the arguments might not be processed in the order they are listed on the command line. This command requires the installation of the optional HP MirrorDisk/UX software (not included in the standard HP-UX operating system) before it can be used. DESCRIPTION
The command splits a mirrored logical volume, lv_path, into two logical volumes. A second logical volume is created containing one copy of the data. The original logical volume is appropriately reset to have one fewer mirror copies. If the option is specified, the new logical volume name has the form lv_pathsuffix. If is not specified, suffix defaults to as in If more than one lv_path is specified on the command line, ensures that all logical volumes are brought offline together in one system call, ensuring predictable results among the logical volumes. All logical volumes must belong to the same volume group. The current num- ber of logical volumes in the volume group added to the number of logical volumes specified on the command line must not exceed the maximum number of logical volumes allowed in the volume group. If PhysicalVolumeGroup is specified, the offline logical volumes are created using the mirror copies on the physical volumes contained in the specified physical volume group. When a mirrored logical volume of a non-shared volume group is split into two logical volumes, a bit map is stored that keeps track of all writes to either logical volume in the split pair. If the volume group is shared, this bit map is not created. When the two logical vol- umes are subsequently merged using the bit map, if present, is used to decide which areas of the logical volumes need to be resynchronized (see lvmerge(1M)). This bit map remains in existence until one of the following conditions occurs: o The merge is completed. o One of the logical volumes is extended, reduced, or split again. o The volume group is cross-activated to shared mode (see vgchange(1M)). o The system is rebooted. The new logical volume must be checked with the command before it is mounted (see fsck(1M)). flushes the file system to a consistent state except for pipes and unlinked but open files. To rejoin two split copies of a logical volume, use the command (see lvmerge(1M)). Options and Arguments recognizes the following options and arguments: lv_path The block device path name of a logical volume. Multiple logical volumes in the same volume group can be speci- fied. The current number of logical volumes in the volume group added to the number of logical volumes speci- fied on the command line must not exceed the maximum number of logical volumes allowed in the volume group. Set automatic backup for invocation of this command. autobackup can have one of the following values: Automatically back up configuration changes made to the logical volume. This is the default. After this command executes, the command (see vgcfgbackup(1M)) is executed for the volume group to which the logical volume belongs. Do not back up configuration changes this time. The offline logical volumes will be created using the mirror copies on the physical volumes in the specified PhysicalVolumeGroup. Specify the suffix to use to identify the new logical volume. The new logical volume name has the form lv_pathsuffix. If is omitted, suffix defaults to as in Shared Volume Group Considerations For volume group version 1.0 and 2.0, cannot be used if the volume group is activated in shared mode. For volume groups version 2.1 (or higher), can be performed when activated in either shared, exclusive, or standalone mode. Note that the daemon must be running on all the nodes sharing a volume group activated in shared mode. See lvmpud(1M). When is issued, it creates the new logical volume device special files on all the nodes sharing the volume group. The device special files are created with the same name on the nodes sharing the volume group. When a node wants to share the volume group, the user must first execute a if logical volumes were split at the time the volume group was not activated on that node. The logical volumes device special files should have the same name on all the nodes sharing the volume group. LVM shared mode is currently only available in Serviceguard clusters. EXTERNAL INFLUENCES
Environment Variables determines the language in which messages are displayed. If is not specified or is null, it defaults to "C" (see lang(5)). If any internationalization variable contains an invalid setting, all internationalization variables default to "C" (see environ(5)). EXAMPLES
Split the mirrored logical volume into two copies. Call the new logical volume Split the mirrored logical volume into two copies. The offline logical volume will be created using the mirror copy on the physical vol- umes contained in the physical volume group Split an online logical volume which is currently mounted on so that a backup can take place: Perform a backup operation, then: Split two logical volumes at the same time: Perform a backup operation on the split logical volumes, then rejoin them: SEE ALSO
lvcreate(1M), lvextend(1M), lvmerge(1M), lvmpud(1M). Requires Optional HP MirrorDisk/UX Software lvsplit(1M)
All times are GMT -4. The time now is 07:26 AM.
Unix & Linux Forums Content Copyright 1993-2022. All Rights Reserved.
Privacy Policy