Description of problem: systemd always marks one swap unit as failed on boot even though it seems to be started ok # systemctl |grep failed dev-sda2.swap loaded failed failed /dev/sda2 # swapon -s Filename Type Size Used Priority /dev/sdb2 partition 2096476 0 0 /dev/sda2 partition 2096476 0 0 # rpm -qa systemd* systemd-sysv-42-1.fc17.x86_64 systemd-analyze-42-1.fc17.x86_64 systemd-gtk-42-1.fc17.x86_64 systemd-42-1.fc17.x86_64
How are the swaps specified in /etc/fstab? Do you see any errors from 'swapon' in /var/log/messages or dmesg?
# systemctl |grep failed dev-sdb2.swap loaded failed failed /dev/sdb2 # dmesg |grep swap [ 29.446565] ALSA sound/pci/hda/patch_realtek.c:1259 SKU: swap=0x0 [ 30.921519] Adding 2096476k swap on /dev/sda2. Priority:0 extents:1 across:2096476k [ 31.044349] Adding 2096476k swap on /dev/sdb2. Priority:0 extents:1 across:2096476k # grep swap /etc/fstab UUID=ec420dbe-b6a9-48d4-ae67-1d3a0310ffb2 swap swap defaults 0 0 UUID=03924499-649c-41b8-9f63-d389a473c02f swap swap defaults 0 0 # grep swap /var/log/messages Feb 12 14:27:46 arekh systemd[1]: Failed to reread /proc/swaps: File exists Feb 12 14:27:46 arekh systemd[1]: Failed to reread /proc/swaps: File exists Feb 12 14:27:46 arekh kernel: [ 30.921519] Adding 2096476k swap on /dev/sda2. Priority:0 extents:1 across:2096476k Feb 12 14:27:46 arekh kernel: [ 31.044349] Adding 2096476k swap on /dev/sdb2. Priority:0 extents:1 across:2096476k Feb 12 19:04:46 arekh swapon[6558]: swapon: /dev/sdb2 : échec de swapon: Périphérique ou ressource occupé Feb 12 19:04:46 arekh systemd[1]: dev-sdb2.swap swap process exited, code=exited status=255 Feb 12 19:04:46 arekh systemd[1]: Unit dev-sdb2.swap entered failed state. Feb 13 03:37:24 arekh swapon[30967]: swapon: /dev/sdb2 : échec de swapon: Périphérique ou ressource occupé Feb 13 03:37:24 arekh systemd[1]: dev-sdb2.swap swap process exited, code=exited status=255 Feb 13 03:37:24 arekh systemd[1]: Unit dev-sdb2.swap entered failed state. Feb 13 19:17:09 arekh swapon[7313]: swapon: /dev/sdb2 : échec de swapon: Périphérique ou ressource occupé Feb 13 19:17:09 arekh systemd[1]: dev-sdb2.swap swap process exited, code=exited status=255 Feb 13 19:17:09 arekh systemd[1]: Unit dev-sdb2.swap entered failed state. Feb 13 19:17:10 arekh swapon[7322]: swapon: /dev/sdb2 : échec de swapon: Périphérique ou ressource occupé Feb 13 19:17:10 arekh systemd[1]: dev-sdb2.swap swap process exited, code=exited status=255 Feb 13 19:17:10 arekh systemd[1]: Unit dev-sdb2.swap entered failed state. Feb 13 19:17:10 arekh swapon[7330]: swapon: /dev/sdb2 : échec de swapon: Périphérique ou ressource occupé Feb 13 19:17:10 arekh systemd[1]: dev-sdb2.swap swap process exited, code=exited status=255 Feb 13 19:17:10 arekh systemd[1]: Unit dev-sdb2.swap entered failed state. Feb 13 19:19:00 arekh swapon[7413]: swapon: /dev/sdb2 : échec de swapon: Périphérique ou ressource occupé Feb 13 19:19:00 arekh systemd[1]: dev-sdb2.swap swap process exited, code=exited status=255 Feb 13 19:19:00 arekh systemd[1]: Unit dev-sdb2.swap entered failed state. Feb 13 19:19:01 arekh swapon[7419]: swapon: /dev/sdb2 : échec de swapon: Périphérique ou ressource occupé Feb 13 19:19:01 arekh systemd[1]: dev-sdb2.swap swap process exited, code=exited status=255 Feb 13 19:19:01 arekh systemd[1]: Unit dev-sdb2.swap entered failed state. Feb 13 19:19:30 arekh swapon[7460]: swapon: /dev/sdb2 : échec de swapon: Périphérique ou ressource occupé Feb 13 19:19:30 arekh systemd[1]: dev-sdb2.swap swap process exited, code=exited status=255 Feb 13 19:19:30 arekh systemd[1]: Unit dev-sdb2.swap entered failed state. Feb 13 19:19:50 arekh swapon[7500]: swapon: /dev/sdb2 : échec de swapon: Périphérique ou ressource occupé Feb 13 19:19:50 arekh systemd[1]: dev-sdb2.swap swap process exited, code=exited status=255 Feb 13 19:19:50 arekh systemd[1]: Unit dev-sdb2.swap entered failed state. Feb 13 19:20:17 arekh swapon[7565]: swapon: /dev/sdb2 : échec de swapon: Périphérique ou ressource occupé Feb 13 19:20:17 arekh systemd[1]: dev-sdb2.swap swap process exited, code=exited status=255 Feb 13 19:20:17 arekh systemd[1]: Unit dev-sdb2.swap entered failed state.
(In reply to comment #2) > Feb 12 14:27:46 arekh systemd[1]: Failed to reread /proc/swaps: File exists The EEXIST is suspicious. It indicates that a device name belonged to one device at one point but later it pointed to a different device. Please attach the output of two commands: udevadm info -q all -n /dev/sda2 udevadm info -q all -n /dev/sdb2
Created attachment 562008 [details] udevadm info -q all -n /dev/sda2
Created attachment 562009 [details] udevadm info -q all -n /dev/sdb2
How come they both have the same symlink?: S: disk/by-path/pci-0000:00:1f.2-scsi-0:0:0:0-part2 Kay?
*** Bug 726019 has been marked as a duplicate of this bug. ***
Hmm, looks like a bug in udev/kernel, reassigning for now. Let's see.
Can you paste /proc/swap when this happens please?
The duplicated by-path/ links for the ATA/SCSI path are caused by a kernel change: "[libata] Add ATA transport class" d9027470b88631d0956ac37cdadfdeb9cdcf2c99 Udev's path_id does at the moment not support the changed sysfs layout for ATA devices with multiple buses on the same host adapter. Identical symlinks claimed by different devices at the same time can always happen, and must not cause any problem with systemd, unless these links are actually used in fstab.
The by-path/ links for ATA transport class devices are now disabled in udev's path_id in the upstream repository, and they will no longer be created; they are not supported until ATA stops faking SCSI devices or someone will add the needed logic to path_id.
Anyway, would still be interested in the contents of /proc/swaps when this happens.
# systemctl |grep failed ddclient.service loaded failed failed LSB: Client to update dynamic DNS host entries dev-sda2.swap loaded failed failed /dev/sda2 # cat /proc/swaps Filename Type Size Used Priority /dev/sdb2 partition 2096476 180292 0 /dev/sda2 partition 2096476 180520 0
Judging by #11 this issue is fixed on newer udev?
it seems to work now indeed
This bug appears to have been reported against 'rawhide' during the Fedora 19 development cycle. Changing version to '19'. (As we did not run this process for some time, it could affect also pre-Fedora 19 development cycle bugs. We are very sorry. It will help us with cleanup during Fedora 19 End Of Life. Thank you.) More information and reason for this action is here: https://fedoraproject.org/wiki/BugZappers/HouseKeeping/Fedora19
Closing as per comment #15.