[sheepdog-users] Cluster snapshot load not working

Liu Yuan namei.unix at gmail.com
Tue Dec 17 16:13:28 CET 2013


On Tue, Dec 17, 2013 at 03:50:25PM +0100, Valerio Pachera wrote:
> I've done htis procedure:
> 
> git pull
> make uninstall
> make clean
> git checkout -b devel origin/devel
> ./autogen.sh && ./configure --enable-zookeeper && make -j 2 install
> 
> sheep -v
>   Sheepdog daemon version 0.7.0_197_g9f718d2
> 
> Starting sheep I get:
> Dec 17 15:23:25   INFO [main] main(887) shutdown
> Dec 17 15:23:25   INFO [main] zk_leave(813) leaving from cluster
> Dec 17 15:28:24   INFO [main] md_add_disk(310) /mnt/sheep/dsk02, vdisk nr
> 233, total disk 1
> Dec 17 15:28:24  ERROR [main] init_config_file(99) This sheep version is
> not compatible with the existing data layout, 2
> Dec 17 15:28:24  ERROR [main] init_config_file(116) use '-u' option to
> upgrade sheep store
> Dec 17 15:28:24  ERROR [main] crash_handler(528) sheep pid 345 exited
> unexpectedly.
> 
> I removed all files and format --copies 2 the cluster.
> 
> It seems to work fine:
> 
> before 'cluster snapshot save'
>   dog vdi read -s 1 wheezy_template | md5sum
>   00e6a6ec009c6218ac37c4aebd8c4766  -
> 
> after 'cluster snapshot load'
>   dog vdi read wheezy_template -s 1 | md5sum
>   00e6a6ec009c6218ac37c4aebd8c4766  -
> 
> I also run the guest before and after, and it works fine.

Thanks for your testing. I'd notice you of that because of this bug, all the
previous 'cluster snapshot save' would be broken because the bug was introduced
before v0.7, so stable-0.7 has this bug from the beginning.

Thanks
Yuan



More information about the sheepdog-users mailing list