FlashBlade is almost ready
May 10, 2020
FlashBlade is almost ready
Engineer in EngineeringTelecommunications Company, 501-1000 employees
Score 7 out of 10
Overall Satisfaction with Pure Storage FlashBlade
We intend to replace our current storage infrastructure for a particular carrier in 2 data centers with FlashBlades due to our requirement for fast NFS storage with encryption at rest. We are still in the process of integrating and are possibly at a point of implementing. Until release 3 came out, the snapshots did not suit our needs and we were unable to integrate it with our 389ds LDAP, but after a bunch of RFEs, it's finally usable. We are still not out of the woods in that the new replication, it doesn't yet meet our needs. On the positive side, Pure is good to work with, my account managers and support engineers have been receptive to my requests and I think their product is improved by this process. The hardware and support are good, but it's very focussed on what it does and needs a better feature set.
- Simple to use
- Non-disruptive upgrades
- Encryption at rest
- NFS version support, v4.1 is not fully featured to the RFC and v4 support would be nice.
- Simple feature set, need to be more enterprise focussed.
- No tool to clear NFS locks.
- No expert mode for troubleshooting no ping, trace route, etc.
- Logs are passworded.
- RBAC roles are limited and not granular enough.
- No local user options other than pure user LDAP/AD only.
- No SSH key trust, so no ansible automation possibilities.
- Replication needs work, can't manage bandwidth without involving support.
- Replication can't mount replica as it only reads normal filesystems, only snapshot view.
- Well, cost wise, the array is priced favorably against the competition.
- On the downside we have burned a lot of man hours on integration.
- You have to buy external software to do things that are built in to other arrays.
We have Oracle ZFS Appliances and the cost of Flashblade was equivalent to one shelf of SSDs in cost. We shopped Isilon and Oracle Solutions, but FlashBlade had a definite cost/performance advantage.
FlashBlade is well suited to performance NFS applications that require encryption at rest, it is good with metadata and scale-out, it would make a good backup target. Integration requires a bit of work since the infrastructure like LDAP servers, etc. are required since the features are external to the device. Compared to my other arrays like NetApp, Oracle, etc., it lacks features that mean I have to log support calls where I could otherwise reasonably troubleshoot or fix the issue myself. I think this is a Pure Storage design choice that has been taken a little bit too far. In my case, this means it has taken a lot longer for me to be productive with the array. We had to rewrite backup scripts to use REST since that was the only option, etc.