| Summary: | sys-fs/lvm2-2.02.109 - Error: /run/lvm/lvmetad.socket: connect failed: No such file or directory | ||
|---|---|---|---|
| Product: | Gentoo Linux | Reporter: | Norman Back <gentoo3> |
| Component: | [OLD] Core system | Assignee: | Robin Johnson <robbat2> |
| Status: | RESOLVED FIXED | ||
| Severity: | normal | CC: | agk, klaus.kreil, ssuominen, weber |
| Priority: | Normal | ||
| Version: | unspecified | ||
| Hardware: | All | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Package list: | Runtime testing required: | --- | |
| Attachments: | rc.log | ||
|
Description
Norman Back
2014-08-24 08:21:14 UTC
The same issue with the startup error message here (using 3.15.5.8-hardened).
It's probably worth mentioning that the socket /run/lvm/lvmetad.socket does actually exist (checked with ls) and there is also an lvmetad process listening on that socket (checked with netstat) on my system *after* booting has finished.
One other interesting observation I can't really make much sense of:
lvmetad, according to an on-screen boot message reading "Starting lvmeatd ...", is started *prior* to lvm (which itself is linked to the the boot runlevel). That would make perfect sense as lvmeatd is listed as a "need" dependency in lvm's init script (NOTE: lvmetad is currently *NOT* explicitly listed in any runlevel on my system).
However, further down the boot sequence after displaying "INIT: Entering runlevel: 3" there's another line reading "Starting lvmeatd ..." - which I don't understand as there's no other dependency that requires lvmetad - so I don't get where that's triggered from.
I do however suspect that only the second incarnation of lvmetad creates the socket and the first one fails for whatever reason - though without issuing an error message. This seems to be supported by a quick test I did: I added
ewaitfile 5 /run/lvm/lvmetad.socket
prior to the following line in /etc/init.d/lvm (to rule out timing issues):
printf "%b\n" "${lvm_commands}" | $lvm_path /proc/self/fd/0 --config ...
A reboot resulted in a 5 seconds delay (the timeout value) before prining the error messages.
To me it seems that the inital start of lvmetad (as a dependency of lvm) fails and/or does not create the socket, whereas the second start (which happens for reasons unknown to me) succeeds and/or creates the socket.
I hope that helps KK
After upgrading to sys-fs/lvm2-2.02.109...... my lvm partitions was not mounted...ufff Ahh, okay i see - i added use_lvmetad =1 to lvm.conf, reboot AHHH all is fine again. BUT: pvdisplay: # pvdisplay No device found for PV tGtthZ-oH7j-V6rc-YP1a-t9aE-4HHF-vrFdAE. vgdisplay: # vgdisplay No device found for PV tGtthZ-oH7j-V6rc-YP1a-t9aE-4HHF-vrFdAE. No volume groups found lvdisplay # lvdisplay No device found for PV tGtthZ-oH7j-V6rc-YP1a-t9aE-4HHF-vrFdAE. No volume groups found tGtthZ-oH7j-V6rc-YP1a-t9aE-4HHF-vrFdAE is the PhysicalVolume. when i fire up blkid: /dev/md3: UUID="tGtthZ-oH7j-V6rc-YP1a-t9aE-4HHF-vrFdAE" TYPE="LVM2_member" So what?!?!?! md3 has the right UUID, why is LVM telling it cant find any device? Is this a bug? Can i secure create new logical volumes? Is data lose possible? (In reply to Marko Weber Bürgermeister from comment #2) > After upgrading to sys-fs/lvm2-2.02.109...... > > my lvm partitions was not mounted...ufff > Ahh, okay i see - i added use_lvmetad =1 to lvm.conf, reboot AHHH > all is fine again 'use_lvmetad = 1' was used as default configuration for systemd users since 2.02.108'ish, but now it's also used for non-systemd users starting with 2.02.110 Does 2.02.110 solve everything you are experiencing? *** Bug 521350 has been marked as a duplicate of this bug. *** (In reply to Samuli Suominen from comment #3) > (In reply to Marko Weber Bürgermeister from comment #2) > > After upgrading to sys-fs/lvm2-2.02.109...... > > > > my lvm partitions was not mounted...ufff > > Ahh, okay i see - i added use_lvmetad =1 to lvm.conf, reboot AHHH > > all is fine again > > 'use_lvmetad = 1' was used as default configuration for systemd users since > 2.02.108'ish, but now it's also used for non-systemd users starting with > 2.02.110 > > Does 2.02.110 solve everything you are experiencing? Hello, all fine with 2.02.110. pvdisplay vgdisplay lvdisplay works now without error. marko I've tried 2.02.110, it is an improvement, in that pvs vgs & lvs are found but I still get an error in /var/log/rc.log * Starting lvmetad ... [ ok ] * Setting up the Logical Volume Manager ... /run/lvm/lvmetad.socket: connect failed: No such file or directory WARNING: Failed to connect to lvmetad. Falling back to internal scanning. [ ok ] Created attachment 383930 [details]
rc.log
|
