Cold Tier on Pure FlashBlade

FlashBlade filesystems

• Always create a separate NFS filesystem for every indexer to host the Cold tier.

• FlashBlade filesystems support both NFSv3 and NFSv4.1 protocols. Choose either one as per your requirement.

• FlashBlade filesystems are always thin-provisioned, Splunk Administrators can provision a large-sized filesystem to avoid updating the size to meet the space growth.

• Do not set the hard limit parameter for the filesystem size as this will limit the flexibility of adding more space as needed.

• Keep all the NFS filesystems for all the indexers in a cluster of the same size to avoid unbalanced space usage.

Linux Mount options

Use the following mount options to mount the NFS filesystem on the indexer nodes for the Cold tier.

rw,bg,nointr,hard,tcp,vers=[3|4.1],rsize=16384

• Select one of the NFS protocol "3" or "4.1" for the vers option.
  • Alternatively, use the nfs4 mount type without the vers option. $ mount -t nfs4 -o rw,bg,hard,nointr,tcp,rsize=16384 10.21.214.203:/splunk-cold01 /cold
• Always mount the filesystem with "hard" mount option and do not use "soft" NFS mounts.
• Do not disable attribute caching.
• Do not specify the wsize option as the host can get the default size offered by FlashBlade (512K).
• To persist these changes across reboots, please include them in the /etc/fstab file as given below. The IP address specified below refers to the data VIP from the FlashBlade.

10.21.214.200:/splunk-cold01 /cold nfs rw,bg,nointr,hard,tcp,vers=3,rsize=16384
or
10.21.214.200:/splunk-cold01 /cold nfs4 rw,bg,nointr,hard,tcp,rsize=16384

Note: Changing the default rsize from 512K to 16K or 32K offers a better read performance.