Hi, I take the chance to mention an issue we have been having since ever since with data protector 9, and that I have been struggling to get acknowledged by support. Our write to tape performance (LTO6) is not optimal because of random freezes of the write operation (it stops writing at intervals). This results in a reduced average speed and we cannot get close to the LTO6 spec for write speed.
Tests performed to confirm it is indeed something that has to do with how DP performs the write operation:
1) Diagnostics over our both our IBM LTO6 drives confirm optimal write speed.
2) Writing to Null device does not present the issue, and shows that the data stream coming from the disk agents is optimal (and generally much faster than the 160MB/s that LTO6 can write at).
3) We can somewhat overcome the issue writing a backup to 2 tape devices at once. This way when one write operation freezes the other drive will be most likely be writing resulting in higher average throughput.
It would be great if we could have the chance to demonstrate this behavior on a webex to some higher tier engineer or at least get an explaination of how this is normal/expected. Of course we are available to collect any debug logs or perform tests
Thanks