Leaderboard

Ok, so my latest write report came in and I am quite happy with the results. Before disabling docker logs / extending helthchecks to 1hour on a BTRFS formatted drive with a BTRFS docker image: 75-85gb writes/day Disabling the "low hanging fruit" docker logs on the same btrfs/btrfs setup: 50gb/day Disabling almost all the docker logs and extending healthchcks to 1 hour 39gb/day Have a feeling that would come down a bit more if left for longer, so basically cut the writes in half, which is about right when you consider the next datapoint. When I first started logging the writes in docker the activity log was around ~100mb/day. Today I got a mere 2.5mb/day, most of that came when one container did some kind of update and flooded the log, excluding that it is like 0.5mb I also started monitoring writes in appdata and the activity log for it is 75mb/day, so all said, cutting writes in half is about right. If I go back to an XFS formatted drive, I should cut the writes by 3x, dropping me down to ~12gb/day, not bad at all. This is as far as I can reasonably take the docker side of the write issue, the remaining logs I do need on occasion so worth a few writes for them. Everything else is program dependent in appdata.

lol, yeah I tend to do that from time to time. Plus seeing the writes slowly add up and the remaining life slowly tick down on the SSD has been an itch I needed to scratch for some time. Figure I will see what is possible under the best case setup and then back off to comfortable compromise after that. For example the above docker side commands are well worth the effort IMHO as they "just work" once you know what containers need which command and have a big effect on writes. Some of the appdata writes are pretty simple fixes as well such as reducing/disabling logging or moving logs to an internal ramdrive instead of writing to appdata etc. My goal is to create another thread with a guide once I finish my testing. For now these posts are basically my journal on the matter for later reference.

For Unraid version 6.10 I have replaced the Docker macvlan driver for the Docker ipvlan driver. IPvlan is a new twist on the tried and true network virtualization technique. The Linux implementations are extremely lightweight because rather than using the traditional Linux bridge for isolation, they are associated to a Linux Ethernet interface or sub-interface to enforce separation between networks and connectivity to the physical network. The end-user doesn't have to do anything special. At startup legacy networks are automatically removed and replaced by the new network approach. Please test once 6.10 becomes available. Internal testing looks very good so far.