Disk Cluster Size

[Peacock]

Hi,

I am having some poor performance with my StarWind VTL library and I am wondering if there are any ways to optimize the disk storage?

Here is the setup:
- ASUSTOR 12R PRO with 12x Seagate Exos 18 TB HDDs, in RAID 6 with ext4 as file system. 4x 1Gb network ports in LACP bond.
- 2x iSCSI targets on ASUSTOR of 32 TB each, Backup server connected to these iSCSI targets.
- Backup server has the 2x iSCSI LUNs mounted and formatted as NTFS with 64k cluster size.
- Backup server is a Windows Server 2019 VM, running StarWind VTL and Veritas Backup Exec.
- StarWind VTL software on Backup server has tape files on the iSCSI mounted drive.
- Backup Exec currently uses 64 KB as tape drive block size, and 512 KB as buffer size.

Currently getting 15 MB/s speed with 100% disk active time according to Windows Resource Monitor.

Must any of these settings be adjusted to result in better performance? Maybe the cluster size of NTFS, or the block size of the tape drive in Backup Exec? Or is there a different/better way of assigning the storage?

Any help would be appreciated.

[yaroslav (staff)]

Greetings,

Could you check the underlying storage performance under 64K stripe size? See more info on benchmarking here https://www.starwindsoftware.com/best-p ... practices/