Comprehensive data protection for all workloads
Post Reply
chiller15
Novice
Posts: 7
Liked: never
Joined: Mar 16, 2017 8:54 am
Contact:

Rebuilding SAN repository

Post by chiller15 »

Due to storage limitations, I have replaced the 9x 3TB drives in RAID5 in my SAN with 5x 8TB drives in RAID6 in our QNAP TVS-1271U-RP. The QNAP is connected via 10GbE to our production storage (HP MSA) and the data is also copied from the QNAP to cloud storage via RClone for archival/disaster recovery purposes. Our new RAID6 array has synchronised and I'm ready to create a storage pool, however I'm not sure which method to use and have some questions:

1) Should I use iSCSI or a standard SMB share?
2) Should I use ReFS or NTFS (our VEEAM is running on a Server 2019 VM - ESXi)
3) Are there any other considerations I need to consider?

Thanks
Egor Yakovlev
Veeam Software
Posts: 2537
Liked: 683 times
Joined: Jun 14, 2013 9:30 am
Full Name: Egor Yakovlev
Location: Prague, Czech Republic
Contact:

Re: Rebuilding SAN repository

Post by Egor Yakovlev » 1 person likes this post

Hi Chiller,

Check out our Best Practices page for Repository Types. That will answer your listed questions.

Thanks!
chiller15
Novice
Posts: 7
Liked: never
Joined: Mar 16, 2017 8:54 am
Contact:

Re: Rebuilding SAN repository

Post by chiller15 »

Hi Egor, thanks for your response. After reading through that documentation, it appears that as this QnAP is a VEEAM backup repository with nightly incremental backups, iSCSI and ReFS is the recommended approach for my use case. However I did notice that the documentation mentions that the SAN requires at least 1GB per TB of used storage and I'll have 21TB of used storage. Does it really make that much of a difference?

I also read somewhere (don't know how accurate is may be) that ReFS still has odd issues on storage that is more than 50% full, even on Server 2019. Is this still the case?
Egor Yakovlev
Veeam Software
Posts: 2537
Liked: 683 times
Joined: Jun 14, 2013 9:30 am
Full Name: Egor Yakovlev
Location: Prague, Czech Republic
Contact:

Re: Rebuilding SAN repository

Post by Egor Yakovlev »

RAM spikes problem was fixed by Microsoft by now, so if your server is up to date, you can easily ignore 1GB RAM\1TB Used rule.
ReFS received a lot of love from community and Microsoft did a huge work fixing found issues. Yes it used to be glitchy, but now its much more stable than before. Feel free to check existing ReFS Problems thread(its quite lenghty, so start from the last page to see up to date user reports)

/Cheers!
Gostev
Chief Product Officer
Posts: 31561
Liked: 6725 times
Joined: Jan 01, 2006 1:01 am
Location: Baar, Switzerland
Contact:

Re: Rebuilding SAN repository

Post by Gostev »

chiller15 wrote: Nov 20, 2019 9:51 amReFS still has odd issues on storage that is more than 50% full, even on Server 2019. Is this still the case?
Yes - but these teething issues are specific to Server 2019 (as it always happens with new Windows releases).
Server 2016 on the other hand is quite solid, and is a good choice until Server 2019 reaches the "SP1" stage ;)
Post Reply

Who is online

Users browsing this forum: Google [Bot], Semrush [Bot] and 115 guests