RepoMirrors/ceph
1
0
Fork 0
mirror of https://github.com/ceph/ceph synced 2024-12-17 08:57:28 +00:00
Code Issues Packages Projects Releases Wiki Activity

mds: ignore ENOENT on writing backtrace

Browse Source 
We get ENOENT when a pool doesn't exist.  This can
happen because we don't prevent people deleting
former cephfs data pools whose files may not have
had their metadata flushed yet.

http://tracker.ceph.com/issues/19401
Signed-off-by: John Spray <john.spray@redhat.com>
This commit is contained in:
John Spray 2017-03-28 14:13:33 -04:00
parent 8442cbb521
commit 3fccc2372f
1 changed files with 17 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

17
src/mds/CInode.cc
View File

@ -1202,6 +1202,23 @@ void CInode::store_backtrace(MDSInternalContextBase *fin, int op_prio)
void CInode::_stored_backtrace(int r, version_t v, Context *fin)
{
if (r == -ENOENT) {
const int64_t pool = get_backtrace_pool();
bool exists = mdcache->mds->objecter->with_osdmap(
[pool](const OSDMap &osd_map) {
return osd_map.have_pg_pool(pool);
});
// This ENOENT is because the pool doesn't exist (the user deleted it
// out from under us), so the backtrace can never be written, so pretend
// to succeed so that the user can proceed to e.g. delete the file.
if (!exists) {
dout(4) << "store_backtrace got ENOENT: a data pool was deleted "
"beneath us!" << dendl;
r = 0;
}
}
if (r < 0) {
dout(1) << "store backtrace error " << r << " v " << v << dendl;
mdcache->mds->clog->error() << "failed to store backtrace on ino "