guestfsd process dead while quitting guestfish
I've compiled guestfsd and run it on CentOS 6.3. It worked well, but when I quit guestfish, the guestfsd process in guest is always dead itself automatically. Is this a bug? or did I miss something? thanks.
On 09/13/2012 05:52 PM, Richard Huang wrote:
I've compiled guestfsd and run it on CentOS 6.3. It worked well, but when I quit guestfish, the guestfsd process in guest is always dead itself automatically. Is this a bug? or did I miss something?
Yeah, the guestfsd should be dead when you finish your operations to the guest, and will be started when you use guestfish next time. It's the design of guestfs, not a bug. Thanks, Wanlong Gao
thanks.
_______________________________________________ Libguestfs mailing list Libguestfs@redhat.com https://www.redhat.com/mailman/listinfo/libguestfs
Since the guestfsd is already dead, that means there is no guestfsd process in the guest, then guestfish cannot execute "run" with --live option any more. In Fedora 17, there is guestfsd.service which uses systemd restart function to bring the guestfsd up I guess.. On Thu, Sep 13, 2012 at 5:57 PM, Wanlong Gao <gaowanlong@cn.fujitsu.com>wrote:
On 09/13/2012 05:52 PM, Richard Huang wrote:
I've compiled guestfsd and run it on CentOS 6.3. It worked well, but when I quit guestfish, the guestfsd process in guest is always dead itself automatically. Is this a bug? or did I miss something?
Yeah, the guestfsd should be dead when you finish your operations to the guest, and will be started when you use guestfish next time. It's the design of guestfs, not a bug.
Thanks, Wanlong Gao
thanks.
_______________________________________________ Libguestfs mailing list Libguestfs@redhat.com https://www.redhat.com/mailman/listinfo/libguestfs
The messages on guest's screen: # ./guestfsd -vr verbose daemon enabled linux commmand line: BOOT_IMAGE=/boot/vmlinuz-3.5.3-1.fc17.i686.PAE root=UUID=88c7691f-a407-45ad-a2b2-c00f0d7c24cc ro rd.md=0 rd.lvm=0 rd.dm=0 SYSFONT=True KEYTABLE=us rd.luks=0 LANG=en_US.UTF-8 rhgb quiet guestfsd: main_loop: new request, len 0x28 fsync /dev/vda fsync /dev/vdb guestfsd: main_loop: proc 282 (internal_autosync) took 0.23 seconds read: unexpected end of file on fd 3 You can see the last line. thanks. On Thu, Sep 13, 2012 at 6:01 PM, Richard Huang <huangyunsong@gmail.com>wrote:
Since the guestfsd is already dead, that means there is no guestfsd process in the guest, then guestfish cannot execute "run" with --live option any more.
In Fedora 17, there is guestfsd.service which uses systemd restart function to bring the guestfsd up I guess..
On Thu, Sep 13, 2012 at 5:57 PM, Wanlong Gao <gaowanlong@cn.fujitsu.com>wrote:
On 09/13/2012 05:52 PM, Richard Huang wrote:
I've compiled guestfsd and run it on CentOS 6.3. It worked well, but when I quit guestfish, the guestfsd process in guest is always dead itself automatically. Is this a bug? or did I miss something?
Yeah, the guestfsd should be dead when you finish your operations to the guest, and will be started when you use guestfish next time. It's the design of guestfs, not a bug.
Thanks, Wanlong Gao
thanks.
_______________________________________________ Libguestfs mailing list Libguestfs@redhat.com https://www.redhat.com/mailman/listinfo/libguestfs
On Thu, Sep 13, 2012 at 05:52:35PM +0800, Richard Huang wrote:
I've compiled guestfsd and run it on CentOS 6.3. It worked well, but when I quit guestfish, the guestfsd process in guest is always dead itself automatically. Is this a bug? or did I miss something?
This is intentional, and also a bug. libguestfs-live currently has a limit of one user at a time. We normally run it from systemd which restarts guestfsd after each user. See: http://pkgs.fedoraproject.org/cgit/libguestfs.git/tree/guestfsd.service The limit of one user at a time makes sense for regular libguestfs, but not so much for libguestfs-live, where we could easily support multiple parallel users. However the reason we don't support that is because we have only one virtio-serial channel with a fixed name (/dev/virtio-ports/org.libguestfs.channel.0). virtio-serial channels are not multiplexed, and they have to be created in advance [AFAIK -- is this still true?] so you would always have a very awkward situation where max N users are supported concurrently (at the moment, N=1). Assuming we stick with virtio-serial for the libguestfs-live case and don't use something more normal like TCP/IP. So that's why this is also a bug. Fixing it isn't too hard TBH, but no one cares enough about this use-case to have sent any patches so far. Rich. -- Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones virt-df lists disk usage of guests without needing to install any software inside the virtual machine. Supports Linux and Windows. http://et.redhat.com/~rjones/virt-df/
On Thu, Sep 13, 2012 at 11:03:26AM +0100, Richard W.M. Jones wrote:
http://pkgs.fedoraproject.org/cgit/libguestfs.git/tree/guestfsd.service
BTW on CentOS you'll need to write a tiny script that does the equivalent of 'while true; do guestfsd [...]; done' Rich. -- Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones virt-df lists disk usage of guests without needing to install any software inside the virtual machine. Supports Linux and Windows. http://et.redhat.com/~rjones/virt-df/
On Thu, Sep 13, Richard W.M. Jones wrote:
Assuming we stick with virtio-serial for the libguestfs-live case and don't use something more normal like TCP/IP. So that's why this is also a bug.
Using TCP/IP would simplify porting to Xen, because there is currently no virtio-serial replacement for a PV guest. Olaf
participants (4)
-
Olaf Hering -
Richard Huang -
Richard W.M. Jones -
Wanlong Gao