An important note for Users of 2M (2048K) block sizes with BRU Server
Recently we have received a number of reports from users witnessing a [W010] Block sequence error being reported for the first file on a 2nd or later volume in a multi-tape archive during the verify pass. Our lab team has examined the myriad potential circumstances and each user reporting the warning is running BRU Server with a 2048K (2M) block size (also known as the bufsize).
In our lab tests, we have uncovered that the issue is with LTO-6 and later drives and relates to a change in the early waring size used by the tape devices. In an effort to resolve a "RECOVER BUFFERED DATA" issue for other software products, the drive manufacturers have caused this issue for applications like BRU that allow the user to utilize the full variable block mode capabilities of the LTO drives.
To prevent this issue, starting with BRU Server 2.0.5, we are locking out the setting of Block Sizes larger than 1024k (the largest block size that we are able to use without losing the carry-over block from a spanned file). This change will not affect the restore of any archives created with a 2M block size prior to this change.
Additionally, it appears that some changes have been made in the last few firmware versions for these drive types that also improves throughput with smaller block sizes. Please re-run a series of test backups on your specific system using approximately 80-100GB of representative files for for system from the drive that you are normally archiving data from. We suggest testing 64k, 128k, 256k, and 512k before selecting 1024k.