Erik, >From the log, it looks like the garbage collection is running fine from a XenServer perspective. There are no exceptions. If you search the log for "trees" you'll see each run and can trace the pid. Do you see any orphaned vdis (i.e. ones which no longer have a reference in CloudStack->Storage) listed from within XenCenter? If so, you can delete those. Any you no longer need from CloudStack should be deleted from within the CloudStack UI at which point I'd expect XenServer storage garbage collection to clean anything up which needs to be cleaned.
You also asked about knowing when the task is done. That's a little bit hard to do since GC is a lazy task. What I do is look in the SMlog for the pid and then grep tail to see when it's completed. -tim On Wed, May 6, 2015 at 2:17 AM, Erik Weber <[email protected]> wrote: > It has now run for about 10 hours, and although it could be unfinished yet, > I'm not overly confident that it'll help. > > As of yet, there's not a single byte freed. > > Here's the output of a minute of SMlog: http://pastebin.com/hPH7YuQU > > -- > Erik > > On Tue, May 5, 2015 at 11:05 PM, Erik Weber <[email protected]> wrote: > > > You're right Tim, > > > > auto-scan is false. > > I've triggered a sr-scan now. Is there any way to notice if it has > > completed or not? > > > > Here's other-config and sm-config if relevant: > > other-config (MRW): trim_last_triggered: 1430856300.47; auto-scan: false > > sm-config (MRO): allocation: thick; use_vhd: true; last-coalesce-error: > > 1430305998; multipathable: true > > > > -- > > Erik > > > > > > On Tue, May 5, 2015 at 10:57 PM, Tim Mackey <[email protected]> wrote: > > > >> If the automatic SR scan isn't happening, garbage collection also won't > >> happen. This is something I ran into very recently in another project. > >> To > >> check if the auto-scan is disabled, issue the following command: > >> > >> xe sr-list uuid=[your uuid] params=all > >> > >> look at the "other-config" and see if there is an "auto-scan: false" > >> entry. If so, then auto-scan isn't going to run. To kick it off > >> manually, > >> issue the following command: > >> > >> xe sr-scan uuid=[your uuid] > >> > >> Depending upon how much stuff is in there, it could take some time (read > >> hours). If there are still things left kicking around after this is all > >> done, look in the SMlog for exceptions and see what they have to say. > >> > >> -tim > >> > >> On Tue, May 5, 2015 at 4:23 PM, Erik Weber <[email protected]> wrote: > >> > >> > Cleanup is enabled. > >> > > >> > No errors in SMlog, > >> > this is the output of vhd-util scan: http://pastebin.com/amBdJSSG > >> > > >> > -- > >> > Erik > >> > > >> > On Tue, May 5, 2015 at 10:10 PM, Somesh Naidu < > [email protected]> > >> > wrote: > >> > > >> > > No, that is not the expectation unless you have storage cleanup > >> disabled > >> > > (storage.cleanup.enabled). > >> > > > >> > > "quite some stale snapshots " - that sounds like leftovers from > >> failed CS > >> > > snapshot operation and/or failing SR scan on XS due to some corrupt > >> VHDs > >> > > interrupting XS GC. Can you check for errors in SMlog on XS relating > >> to > >> > GC > >> > > on the particular SR? > >> > > > >> > > Somesh > >> > > CloudPlatform Escalations > >> > > Citrix Systems, Inc. > >> > > > >> > > -----Original Message----- > >> > > From: Erik Weber [mailto:[email protected]] > >> > > Sent: Tuesday, May 05, 2015 3:38 PM > >> > > To: [email protected] > >> > > Subject: Rogue / stale snapshots with XenServer > >> > > > >> > > I seem to have quite some stale snapshots on my primary storage, of > >> both > >> > > deleted vms, volumes and snapshots from ACS. > >> > > I'm using FC with PreSetup on XenServer 6.5, with latest RC of ACS > >> 4.5.1 > >> > > > >> > > Is it expected to do manual cleanup on the SR? > >> > > > >> > > -- > >> > > > >> > > Erik > >> > > > >> > > >> > > > > >
