Hi Lennert, I recall having seen such an issue. Is it similar to https://issues.apache.org/jira/browse/CLOUDSTACK-643 ?
Thanks, -Nitin On 06/09/13 5:35 PM, "Lennert den Teuling" <lenn...@pcextreme.nl> wrote: >Hi, > >I was wondering if somebody is experiencing the same issue. Yesterday we >found out that the disks of our secondary storage was almost full >because none of the snapshots that have been deleted (because of >rotation for example). > >The issue seems quite clear (?), cause the commando executed on the SSVM >doesn't work and the SSVMs tells the manager the command was executed >successful which it was not. > > From /var/log/cloud/systemvm.log on the SSVM: >Executing: /bin/bash -c rm -f >/mnt/SecStorage/0bab1727-6381-3daa-980e-8a51fdcbb7cd:snapshots:5:1032:*/sn >apshots/1/5/1032/lnx-lennert-eu_ROOT-844_20130901160036* > > >It looks like the path is incorrect, see the ":". And is specified >twice. If you would remove ":snapshots:5:1032:*" the path is correct. > >Also the manpages of rm for the "-f" flag states: >"deletes read-only files immediately, without asking for confirmation. >When you specify this option and a file does not exist, rm does not >display a warning message and does not modify the exit status. If you >specify both -f and -i, rm uses the option that appears last on the >command line." > >Maybe this could be the reason why the SSVM returns success. > >I did search for a bug report, but could not find one. Maybe i didn't >look good enough or this issue isn't known. > >We are using CS 4.11 with KVM on Ubuntu. Maybe somebody can dig up the >logs on their system and see if they also see these (wrong) rm commands >show up in the log? > >-- >Met vriendelijke groet / Kind regards, > >Lennert den Teuling >
