How IBM Tivoli Storage Manager works with Microsoft VSS

vsscomponentsThis is a great article from IBM on how TSM operates with Microsoft Volume Shadow Copy Service (VSS):

http://www.ibm.com/developerworks/tivoli/library/t-tsm-vss/index.html

Advertisements

An economical DR plan – a case study

network2This is a review of a disaster recovery case study I did in an enterprise storage class in grad school:

I have found a case study of an economical DR plan that contains a bi-coastal partnership between two universities. Bowdoin College in Brunswick Maine had partnered with Loyola Marymount University in Los Angeles and created a plan to host one another’s DR sites. The recovery sites are identical and linked with a high speed VPN. Each recovery site is an active asynchronous secondary site: “If a disaster or outage hits either school, the hosting campus will initialize the other’s hot site and run it for the duration of the emergency” (Cox, 2007). The secondary sites will be operated for the duration of the disaster. Each site will maintain the recovery systems that have been purchased by the other. During a disaster, the administrators would remotely administer the recovery site – a logical data center. Such systems as emergency web sites would be in operation continuously, as well as some DNS elements and a program that distributes an emergency notification through various content delivery systems such as pager, IM, cell phone, and email.
According to the CIO’s at each location have discussed, it would be more affordable to do this within similar organizational environments. To keep cost down, virtualization was implemented at the start – less hardware was needed which also will conserve energy cost. With two organizations working together each could share expertise which was mutually beneficial, and two organizations with similar infrastructure would also be beneficial to the initial design of the recovery site architecture, further reducing cost. Another caveat is that the collaboration between organizations would produce an ‘informal help desk where managers and staff can reach out for advice, brainstorm and troubleshoot problems with their colleagues a continent away.”I can send an instant message to Dan and say ‘have you seen this [problem]?’ and he IM’s back ‘we saw this three weeks ago and here’s what we did.. .’” (Cox, 2007).
Most vital to this methodology is relationship-building between the two organizations, as your data is moving off campus and into another organization.

Reference:
Cox, John. “The case of the great hot-site swap; Colleges create a bicoastal disaster-recovery plan by hosting each others hot sites.(Loyola Marymount University in Los Angeles and Bowdoin College in Brunswick, Maine). .” Network World. (August 3, 2007): 1. Academic OneFile. Gale. BCR Regis University. 14 May 2009

Some backup basics

Incremental: Backs up only data that has changed since the last backup. The least amount of data is backed up with this method, shortening backup windows. This is a quick method that skips any file that has the same date as the same file within the backup archives.
Differential: Back up of all files changed since the last full backup. Upon restore, the full and last differential backup medium will be used. A negative about this method is that throughout the week, the backup will become larger until the next full backup. This can impinge on the backup windows.
CDP – Continuous Data Protection: This is like a snapshot backup – a pint-in-time snapshot of a file system. Oracle has it: RMAN and Flashback – a database snapshot utility that is a set of SQL commands that lets users view data as it existed at various points in time. This enables one to identify points of corruption and to restore a database to a point before the corruption” (Storage Magazine, 2005). IBM Tivoli has it: the client software has the option of point-in-time restore. EMC’s Celerra NAS appliance has the checkpoint utility which takes a complete snapshot of a file system in seconds. NDMP is often used to back up the Celerra NAS checkpoints:
NDMP: Network Data Management Protocol is a popular way to backup immense data on NAS devices. Often a fast LAN-free method, NDMP is efficient in the way data is transferred from the NAS directly to the backup server without any third-party backup client software installed on the NAS device. Using a LAN-free NDMP backup, the LAN is free of traffic from the large NAS backup because NDMP directly accesses the backup server often through SCSI or fiber medium, bypassing the rest of the LAN that contains the other server nodes.
VTL: The latest backup medium, the Virtual Tape Library, is a disk library dedicated to backup data storage. Disk is faster than tape and this method shortens backup windows considerably on a fast fiber-based SAN. The backup database is then dumped to tape periodically and the tape is then transferred to offsite vaulting for DR.

References:
Storage Magazine. (2005). CDP poised to replace traditional backup methods. SearchStorage.com. Retrieved June 3, 2009 from http://searchstorage.techtarget.com/magazineFeature/0,296894,sid5_gci1257934,00.html

Backup and restore a Brocade 4900 series fiber switch

brocadeCreating a back up of a configuration file
Keep a backup copy of the configuration file in case the configuration
is lost or unintentional changes are made. You should keep
individual backup files for all switches in the fabric. You should
avoid copying configurations from one switch to another.
To back up a configuration file
1. Open the Switch Administration window as described on
page 59.
2. Click the Configure tab.
3. Click the Upload/Download subtab (see Figure 28 on page 102).
4. Click the Config Upload radio button.
5. Choose whether the download source is located on the network
or a USB device.
• If you select the USB radio button, you can specify a firmware
path. The USB radio button is available if the USB is present
on the switch.
• If you selected the network as the configuration file source,
type the host IP, user name, file name, and password.
You can enter the IP address in either IPv4 or IPv6 format.
6. Type the configuration file with a fully qualified path.
7. Select a protocol to use to transfer the file.
8. Click Apply.
You can monitor the progress by looking at the
Upload/Download progress bar.

Restoring a configuration
Restoring a configuration involves overwriting the configuration on
the switch by downloading a previously saved backup configuration
file. Perform this procedure during a planned downtime.
Make sure that the configuration file you are downloading is
compatible with your switch model, because configuration files from
other model switches might cause your switch to fail.
Maintaining configurations 105
Maintaining Configurations and Firmware
To download a configuration to the switch
1. Open the Switch Administration window as described on
page 59.
2. Disable the switch, as described in “Enabling and disabling a
switch” on page 71.
You can download configurations only to a disabled (offline)
switch. You will only be able to disable the switch if you the
Admin Domain you are logged into owns the switch.
3. Click the Configure tab.
4. Click the Upload/Download subtab (see Figure 28 on page 102).
5. Click the Config Download to Switch radio button.
6. Choose whether the download source is located on the network
or a USB device.
• If you select the USB radio button, you can specify a firmware
path. The USB radio button is available if the USB is present
on the switch.
• If you selected the network as the configuration file source,
type the host IP, user name, file name, and password.
You can enter the IP address in either IPv4 or IPv6 format.
7. Type the configuration file with a fully qualified path.
8. Select a protocol to use to transfer the file.
9. Click Apply.
You can monitor the progress by looking at the
Upload/Download progress bar.

Restarting the Tivoli AIX client

iscbanner-mosaicThere is an AIX box at work that I have a TSM client on, and I have had to make some changes to it. As I was not experienced in AIX, I had to find how to stop the Tivoli client process and then re-enable it for the changes to take effect. Here is how to do it:

To list the TSM processes, use:
ps -eaf|grep dsmc

To kill them, use:
kill -9 (process id)

Then, restart:
dsmcad
nohup dsmc schedule 2> /dev/null &

…This will start the dsm client access daemon for the web services, and start the scheduler so that it will continue to run, not unlike a service after the administrator logs off of the AIX system.

Useful TSM commands for storage management

iscbanner-mosaicI have been having the issue of my backup client logs reporting “insufficient storage space” when running daily backups. The VTL that Tivoli Storage Manager is using for its primary storage consists of an EMC DL4100 system with 26Tb storage – no way I could be out of space yet. I have found, with some suggestions from IBM is to use this first step to resolution: move data from some volumes that are registed as full but with very little utilization, to otehr volumes to free up space. For example, volume v00021 is registed full, but with only 9% utilization… Why Tivoli is doing this, I have no idea; but one can move the data on that volume to another volume within the same storage pool to free up that space, which converts v00021 to scratch status – enabling free space. This is what to do. For finding volumes with low utilization (say, 0-40%), use:

q vol st=full stg=(your storage pool name)

Then, to move data from those volumes, use:

move data (volumename)

This will move data from that volume to another within the same poll and free up the space. This will help, but make sure that expiration and reclaimation are running in order to utilize that space. To check if any scratch volumes are appearing, enter –

q libv

…Now to find the cause of the storage issue in the first place… -S.J.