A few misc Q's.........

[kiriak]

what do you mean?
I verify my backup tasks every 4-6 months, depending on the task and how frequently their data change.
Each backup task is about 400-600 Gb and the verification of each one takes about 2 hours (or probably more) on a DS 116.
Now that I own a DS218+, I haven't run a verification yet.

 

[Rusty]

Now that one I didn't know

A thought from your earlier post..... how do you go about verifying backups?

Random restore tests. Thats how I do it. Local and remote

 

[kiriak]

There is an option in HyperBackup that verifies the selected backup task.
I run this every 4-6 months.

Besides this, from time to time I check if I can read the backup's contents with the HyperBackup Explorer in both my Windows and Fedora installs.

 

[Coop777]

With regard to Jumbo Frames and MTU - be careful changing those settings, you may find that network performance actually gets worse if you do. On a 1 GBE network stick with defaults. You should notice transfer speeds ~110 mpbps, which is about as good as it will get.

 

[jeyare]

With regard to Jumbo Frames and MTU - be careful changing those settings, you may find that network performance actually gets worse if you do. On a 1 GBE network stick with defaults. You should notice transfer speeds ~110 mpbps, which is about as good as it will get.

Pls. be specific with your advice, because reader can’t understand what was targeted in your post (NAS NIC, network appliances or computer NIC).

LAN MTU setup depends on the LAN architecture and all the rest of setup in E2E traffic (mentioned in my post), then is the most important point in my post:
From my perspective is necessary test all possible scenarios E2E (NAS - computer) to get best performance on each sides.
There is tuning of the setup necessary for best performance achievement. Also with the MTU. And each tuning step must be performed in right way:
- hypothesis
- check
- true/false. And documented evaluation for each future change in your network topology.
Otherwise it isn’t tuning but just shoot in the dark, many times this is the reason of slow speed of your LAN transport.

And there is also different WAN MTU setup based on transport layer (FTTX, xDSL, WI-FI, ...) and transport elements. Then you have to take into account e.g. your Router setup (specially when you have some router in bridge mode (front of your router). Absolutely yes, here you can’t change ISP MTU setup, because your speed will drop significantly.

 

[jeyare]

Right, understood......

However it looks like the GS108 is not managed at all (it is not assigned an IP address) and therefor does not have a web console that can be accessed. It has sped up file transfer between the Mac and the NAS though

some Netgerar unmanaged switches have predefined MTU for 9000, then they forward all the traffic but with large fragmentation loses. It's not clear from Netgear knowledge base which one, but I have seen many times such test results in past (>3y ago).
reason why I'm using all the switches managed - price difference is not high - result is huge. Also small 5-ports.

 

[Rusty]

Running a few managed and unmanaged switches from NETGEAR (3x8port and 1x24port) as well. 0 problems with default jumbo.

 

[tekguru]

Random restore tests. Thats how I do it. Local and remote

Yep I've been doing that I'm glad to say

There is an option in HyperBackup that verifies the selected backup task.
I run this every 4-6 months.

You mean the Backup Integrity Check scheduling within the HyperBackup package on the NAS? Yep turned on

Sounds like I'm good

 

[Coop777]

Pls. be specific with your advice, because reader can’t understand what was targeted in your post (NAS NIC, network appliances or computer NIC).

LAN MTU setup depends on the LAN architecture and all the rest of setup in E2E traffic (mentioned in my post), then is the most important point in my post:
From my perspective is necessary test all possible scenarios E2E (NAS - computer) to get best performance on each sides.
There is tuning of the setup necessary for best performance achievement. Also with the MTU. And each tuning step must be performed in right way:
- hypothesis
- check
- true/false. And documented evaluation for each future change in your network topology.
Otherwise it isn’t tuning but just shoot in the dark, many times this is the reason of slow speed of your LAN transport.

And there is also different WAN MTU setup based on transport layer (FTTX, xDSL, WI-FI, ...) and transport elements. Then you have to take into account e.g. your Router setup (specially when you have some router in bridge mode (front of your router). Absolutely yes, here you can’t change ISP MTU setup, because your speed will drop significantly.

I saw no need to write a book on it nor go into a really deep dive, just giving my .02 on what I have seen.

 

[jeyare]

I saw no need to write a book on it nor go into a really deep dive, just giving my .02 on what I have seen.

look, understand your point and I like most of your responses
I would like to keep an environment, when a help or advice to other person must be better than:
don't use the car because you can crash

 

[Coop777]

no worries - LACP/LAG/Jumbo Frames/MTU is something I spent much of a full weekend implementing in an attempt to increase my NAS bandwidth on a 1 GB network and it didn't help anything. Hoping to spare others the time and frustration.