Hard disk access time, according to size

What difference in file access time can there be between small, medium or large hard drives when they are full?

What difference can there be between 4, 8, 14 or 20Tb hard drives?

I have two nodes.
One node with a 14Tb hard drive with 8.9Tb of storj data saved.
A second node with a 3Tb hard drive with 0.8Tb of storj data saved.

A few weeks ago the big node stopped a few times due to slow read or write access times. The problem seems to be solved by raising the time 30 seconds for reading and writing. I have to wait a few more days to be sure that the access problem is resolved.

I am worried that the node will fail due to slow speed when the 14TB hard drive is full.

I have two 8Tb hard drives that I could change to use for the 14T and 3Tb drive nodes that I have now.

Would it be a good idea to change the nodes to the two 8Tb hard drives?

The first one is too big, you can’t move it to the 8Tb disk

I know that the first node is too big.

I would have to reduce the size of the node and wait a few months for data to be deleted.

Access times on mechanical hard drives are affected by disk rotation speed, platter radius, and read/write head speed. Assuming your disks are all the same form factor and were made within a couple of years of each other, I wouldn’t expect you to see any variation in access times as a function of storage capacity.

I would instead recommend the following:

  1. Check your disk health by looking at it’s smart attributes.
  2. If your drive appears to be healthy, look to see how fragmented your filesystem is, and defragment if applicable.
1 Like

You might be the not so proud owner of an SMR drive…

as @aad stated the speed of a disk is down to mechanics, the size in this case is practically irrelevant…

SMR drives will rewrite multiple tracks at once, thus it will have a vastly slower write speed under continual load.

while a conventional drive writes only one track at a time.

basically the problem is that an SMR drive has to read the other tracks store them in its buffer than then rewrite multiple tracks when it adds new data to a section.

SMR drives are slightly cheaper than CMR drives… but not so much that it matters, ofc for those that only use them for storing movies or mainly do read operations it won’t matter…

but SMR is terrible at sustained writes…
however SMR will read at the same speeds as a CMR drive.

1 Like

i agree that SMR is bad, but most 14TB drives on the market are CMR (WD and Seagate).

out of curiosity, what OS is the node, and what file system on the drives?

my drives are only about 2 or 3TB full, and utilization occassionally spikes to 100% but is usually lower. That’s running linux and btrfs (I turned of btrfs copy on write with chatter +C, that was getting extremely slow on a drive)

Perhaps you would need to migrate from BTRFS ASAP:

I have windows 11 and ntfs file system.

Usually only fragmentation and/or type of the HDD (SMR are incredible slow on writes and could be slow or reads too) is affecting the access time, if you use the native filesystem for your OS (NTFS and Windows, ext4 for Linux).
But it’s possible that some disks could be slower on bigger models (usually not).