After installing an ostree-based desktop distribution of Fedora (Silverblue, Kinoite, etc.), the computer will become slower and slower over time when IO happens, as discard is not properly enabled. The IO problems eventually become so bad that the computer can hang for minutes with no response. It's probably an issue with the way Anaconda sets up the system. It may also be an issue with ostree deployment. (It could be both of these things.) Reproducible: Always Steps to Reproduce: 1. Install Silverblue on a device that needs trim 2. Check to see if you can trim the device (fstrim -va is one method) Actual Results: No discard support. Computer becomes slower and slower on every disk access, which becomes very noticeable after a few months. Expected Results: Discard should happen automatically with fstrim.timer (and when run manually). Workstation is fine. This is a problem only with ostree-based distributions (not inherent to ostree itself, but with the way they're installed).
I solved it locally with `sudo cryptsetup --allow-discards --persistent refresh /dev/mapper/luks*` Vojtech Trefny says `rd.luks.options=discard` via Anaconda should solve it for everyone (on fresh installs).
(In reply to Garrett LeSage from comment #1) > Vojtech Trefny says `rd.luks.options=discard` via Anaconda should solve it > for everyone (on fresh installs). Note this is based on this[1] Fedora Silverblue issue, I am certainly not expert on OSTree/Silverblue and I'm open to different solution. [1] https://github.com/fedora-silverblue/issue-tracker/issues/31
This bug appears to have been reported against 'rawhide' during the Fedora Linux 39 development cycle. Changing version to 39.