Spectator47

joined 1 year ago
[–] Spectator47 2 points 1 year ago

Check out https://www.fs.com/c/25g-smb-switches-4247

FS have inexpensive switching and SFPs. Even if you get another brand switch I still recommend FS SFPs which can be ordered for compatibility with various vendor equipment.

[–] Spectator47 1 points 1 year ago

Yes the one I linked was 100W / 20V / 5A

[–] Spectator47 6 points 1 year ago (2 children)

For power delivery you need a cable rated for it. Cables can have the same USB specification but different power limits.

Typically to charge a Laptop you want a cable rated for 60W power delivery or higher. I would just go with 100W rated cable as the price is reasonable.

Check out https://www.amazon.com/UGREEN-Cable-100W-Fast-Charge/dp/B07V5MKXJ1/ref=mp_s_a_1_3?crid=3G8AD8YL4Q5XE&keywords=ugreen%2B65w%2Busb%2Bc%2Bcable&qid=1692690249&sprefix=ugreen%2Busb%2Bc%2B65w%2Caps%2C370&sr=8-3&th=1&psc=1

[–] Spectator47 1 points 1 year ago

The recordsize used is dynamic up to a maximum of the zfs recordsize property. See https://klarasystems.com/articles/tuning-recordsize-in-openzfs/

The purpose of tuning it is about optimising reads and writes within files larger than the recordsize. For example if you have a database that typically stores its data in a single large file then because it is a large file zfs will be reading and writing it in recordsize chunks on disk. If your databse operates on 4k size changes then reading and writing 1MB at a time in disk is a waste of I/O bandwidth.

[–] Spectator47 1 points 1 year ago

The recordsize used is dynamic up to a maximum of the zfs recordsize property. See https://klarasystems.com/articles/tuning-recordsize-in-openzfs/

The purpose of tuning it is about optimising reads and writes within files larger than the recordsize. For example if you have a database that typically stores its data in a single large file then because it is a large file zfs will be reading and writing it in recordsize chunks on disk. If your databse operates on 4k size changes then reading and writing 1MB at a time in disk is a waste of I/O bandwidth.