Slow RAID0 to RAID0 on 314s with bonded gigabit

mdgm wrote:
I probably would have gone for the 516 for this use case. Furthermore with a single client you would like only saturate one link.

True.

Though if the business requires maximum network performance (and can justify the cost), perhaps even a 716 and 10 gigabit throughout, using RAID-10.

Cheers for the reply. I've been away on holiday for a couple of weeks, so I didn't get to respond. I also wasn't getting much opportunity to run tests before I left.

StephenB wrote:
I think NFS is the best protocol to test with for NAS-to-NAS (why use a Windows file sharing protocol between two Linux machines???)

Because I was working this out as I went via trial and error :wink: I also had limited opportunities for such trial at the time.

At any rate, just before I left I tried using RSync for the backup between the NAS boxes and was able to get over 100MBps throughput, so SMB was certainly the limiting factor there. Syncing is also what I was after in the first place, rather than full/incremental backups. A single gigabit connection is also fine for nightly backups.

StephenB wrote:

One challenge with NIC bonding is that the server also needs to communicate with clients that only have a single NIC. When you use two NICs for the same traffic flow, packets will arrive out-of-order at the receiver, and you can also end up with packet drops in the switch that serves the receiving device. The simplest way to avoid these issues is to send the entire flow for a connection out a single NIC. That's what LACP does (and most of the other non-standardized Linux bonding modes as well).

That makes sense, thanks. I carried out simultaneous read and write tests to/from two computers at a time and both were able to more or less saturate their link, so the current bonding clearly works in that regard, which is definitely the most important goal.

StephenB wrote:
As a practical matter - if you had a NAS failure, wouldn't you immediately switch over to the backup NAS? Why is the speed of restoring the original NAS a major concern?

Mostly because I want to minimise the amount of time the second NAS is the only copy of valuable data, especially since it will probably be working hard at the time. If a single drive in that one dies during that time (which could potentially be a period of multiple days) then the business is pretty much screwed. RAID10 is not currently an affordable option and RAID5/6 seem almost pointless these days. I think I'll make it protocol that should a NAS fail then the most critical current files should be manually copied to external hard drives via the USB3.0 connections straight away.

StephenB wrote:
If you configure bonding to get > 1 gbit flow for a single connection, you will almost certainly see performance issues with your normal single nic clients.

That's fair enough. I've determined that it won't be important during normal operation anyway. It's only in the event of a restore that utilising full bonding speeds to a single device (the other NAS) would be a concern.

Given that in such a situation only tx bonding would be needed on the backup NAS, and only rx bonding on the primary NAS, is there an effective and efficient way to configure the setup to utilise both gigabit connections during the restore (preferably without an enterprise-level priced switch)?

StephenB wrote:
Though if the business requires maximum network performance (and can justify the cost), perhaps even a 716 and 10 gigabit throughout, using RAID-10.

The current setup is definitely the most the business could currently afford :) It's quite young and small. Video is just such a demanding resource even at that level.