I have been designing, installing , and operating backup systems for the past several years. I have mostly implemented and managed Symantec Netbackup (used to be Veritas Netbackup) for larger infrastructures and Symantec Backup Exec for smaller ones.
These software worked very well although some features are not very robust. I’m very impressed for instance of the NDMP implementation in Netbackup. Backing up terabytes of NetApp data via NDMP works very well. However, I do not like the admin user interface of Netbackup since its not very intuitive. Their bare metal restore (BMR) implementation also is a pain. Some of the bugs took years to fix. Maybe because there are not too many companies using BMR.
Backup Exec works very well with small to medium systems. It has very intuitive interface, it is relatively easy to setup, and it has very good troubleshooting tools. Lately though, Symantec has been playing catch up in their support for newer technologies such as VMware. It is so much easier to use Veeam to manage backup and restore of virtual machines. In addition, Backup Exec has been breaking lately. Recent Microsoft patches have caused backup of System_State to hang.
But I think the biggest threat to these backup software are online backup providers. Crashplan, for instance, was initially developed for desktop backup, but it will not take long before companies will use it to back up their servers. When security concerns are addressed properly by these providers, companies will be more compelled to backup their data online. It’s just cheaper and easier to backup online.
One of my latest projects is to consolidate six old NetApp Filers and migrate a total of 30 TB of data to a new NetApp Filer cluster, FAS 3240C. The project started several months ago and it is almost complete. Only one out of six NetApp filers is left to migrate.
I have done several storage migrations in the past, and there are always new lessons to learn in terms of the technology, migration strategy and processes, and the people involved in the project. Here are some of the lessons that I learned:
- As expected, innovations in computer technology move too fast and storage technology is one of them. IT professionals need to keep pace or our skills become irrelevant. I learned storage virtualization, NetApp fast cache, and snapmirror using smtape, among many other new features.
- Migration strategy, planning, and preparation take more time than the actual migration itself. For instance, one filer only took an hour and a half to migrate. However, the preparations such as snapmirroring, re-creating NFS and CIFS shares, making changes in users login scripts, making changes in several applications, and many other pre-work were done several days before the actual migration. The actual migration is actually just to catch up with the latest changes in the files (ie snapmirror update), and flipping the switch.
- People, like many other big IT projects, are always the challenging part. The key is to engage the stakeholders (business users, application owners, technical folks) early on in the project. Communicate with them the changes that are happening and how their applications and accesses to their data will be affected. Give them time to understand and make changes to their applications. Tell them the benefits of the new technology and communicate often the status of the project.