I’m dualbooting Fedora Kinoite (ublue-nvidia image) with Windows 11 and I have a boot time of over 1 minute (only on the Fedora side).
The output of systemd-analyze critical-chain is:
└─sddm.service @16.435s
└─plymouth-quit.service @16.315s +107ms
└─systemd-user-sessions.service @16.299s +12ms
└─remote-fs.target @16.298s
└─remote-fs-pre.target @16.298s
└─nfs-client.target @16.298s
└─gssproxy.service @16.288s +9ms
└─network.target @16.285s
└─wpa_supplicant.service @16.281s +4ms
└─basic.target @14.798s
└─dbus-broker.service @14.774s +22ms
└─dbus.socket @14.760s
└─sysinit.target @14.757s
└─systemd-resolved.service @14.696s +61ms
└─systemd-tmpfiles-setup.service @14.584s +96ms
└─local-fs.target @14.569s
└─run-user-1000-doc.mount @23.123s
└─run-user-1000.mount @22.463s
└─swap.target @1.410s
└─dev-disk-by\x2duuid-819f25f8\x2daf77\x2d4d7b\x2daaf7\x2dadb07819a7b1.swap @1.276s +35ms
└─dev-disk-by\x2duuid-819f25f8\x2daf77\x2d4d7b\x2daaf7\x2dadb07819a7b1.device @584542y 2w 2d 20h 46.792s +1min 3.997s
First of all, I would like to know what the hell is going on with that 584542 years active time lol
Anyway, the x2dadb07819a7b1 UUID belongs to the swap partition.
Output of lsblk -f:
NAME FSTYPE FSVER LABEL UUID FSAVAIL FSUSE% MOUNTPOINTS
zram0 [SWAP]
nvme0n1
├─nvme0n1p1 vfat FAT32 EFI AAFB-90EA 553.6M 7% /boot/efi
├─nvme0n1p2 ext4 1.0 fedora-boot a1457f7b-c1fb-40da-9c6f-98356d9003e2 526.8M 39% /boot
├─nvme0n1p3 ext4 1.0 fedora-root 0e748e63-f5f5-42f1-babd-818054eb9ee5 40.8G 35% /var
│ /sysroot/ostree/deploy/fedora/var
│ /usr
│ /etc
│ /
│ /sysroot
├─nvme0n1p4 swap 1 fedora-swap 819f25f8-af77-4d7b-aaf7-adb07819a7b1 [SWAP]
├─nvme0n1p5 crypto_LUKS 2 ea073ead-906c-4127-9555-efba204baabf
│ └─luks-ea073ead-906c-4127-9555-efba204baabf ext4 1.0 fedora-home e37f299a-84f5-46ce-976c-507b8e8e25f8 1T 1% /var/home
├─nvme0n1p6 ntfs Extra 74FE8F25FE8EDF2C
├─nvme0n1p7
├─nvme0n1p8 BitLocker 2
└─nvme0n1p9 ntfs C02807922807869E
What should I do?
Just speculation on my end but it could be because you’ve got an encrypted home partition on a different partition of the same physical device. If you’ve got enough RAM and ZRAM for swap, why not try disabling the swap partition entirely?
Second this because:
- Ideally, a swap partition should always be on a secondary physical device than your home directory. Or anything else, if possible, as there will be high i/o during times of high memory activity and memory/swap maintenance.
- Your bottleneck is exacerbated more by the excessive overhead of on-the-fly decryption activity on the same physical device
- All of this can be avoided by eliminating swap, which isn’t necessary if you have sufficient RAM.
Exactly. That’s the same theoretical underpinning I had for my guess.
Thanks for laying it out so concisely.
Yeah, I might actually end up disabling swap in the end. I wanted to update that apparently I “fixed” the problem (not sure if permanently) by turning off the pc, unplugging the PSU, and holding down the power button for 30 seconds. Normal reboots weren’t enough. I’ll take it for now.
All of this can be avoided by eliminating swap, which isn’t necessary if you have sufficient RAM
This isn’t true. Swap also exists to offload old cached memory pages. No Swap means your system can’t do that
This explains it a lot better than I could
