From: Sage Weil <sage@redhat.com>
Date: Fri, 12 Aug 2016 19:32:34 +0000 (-0400)
Subject: os/bluestore/BlueFS: assert we never extend log at flush
X-Git-Tag: ses5-milestone5~93^2~6
X-Git-Url: http://git-server-git.apps.pok.os.sepia.ceph.com/?a=commitdiff_plain;h=d0934300b1afb120d555ce6756c9eafdec5ab401;p=ceph.git

os/bluestore/BlueFS: assert we never extend log at flush

Signed-off-by: Sage Weil <sage@redhat.com>
---

diff --git a/src/os/bluestore/BlueFS.cc b/src/os/bluestore/BlueFS.cc
index 394a31d64d82..835b76063fd7 100644
--- a/src/os/bluestore/BlueFS.cc
+++ b/src/os/bluestore/BlueFS.cc
@@ -1210,6 +1210,9 @@ int BlueFS::_flush_range(FileWriter *h, uint64_t offset, uint64_t length)
   // previously allocated extents.
   bool must_dirty = false;
   if (allocated < offset + length) {
+    // we should never run out of log space here; see the min runway check
+    // in _flush_and_sync_log.
+    assert(h->file->fnode.ino != 1);
     int r = _allocate(h->file->fnode.prefer_bdev,
 		      offset + length - allocated,
 		      &h->file->fnode.extents);