replication stops because of a bug in fdsize

[GoogleCodeExporter]

will first found this when he was running zumastor replication. Our cbtb
replication tests also failed because of this bug. The bug was introduced
in revision 1579. Here is the fix.

Index: ddsnap.c
===================================================================
--- ddsnap.c    (revision 1584)
+++ ddsnap.c    (working copy)
@@ -1163,11 +1163,11 @@
        u64 extent_addr = 0, chunk_num;
        int current_time, last_update = 0;

-       if (!fullvolume && (source_volume_size = fdsize64(snapdev1)) != -1) {
+       if (!fullvolume && (source_volume_size = fdsize64(snapdev1)) == -1) {
                warn("unable to determine volume size for %s", dev1name);
                goto out;
        }
-       if ((target_volume_size = fdsize64(snapdev2)) != -1) {
+       if ((target_volume_size = fdsize64(snapdev2)) == -1) {
                warn("unable to determine volume size for %s", dev2name);
                goto out;
        }

Jiaying

Original issue reported on code.google.com by jiahotc...@gmail.com on 25 Apr 2008 at 2:48

[GoogleCodeExporter]

Would the smoke test have found this, or is it still borken?

Original comment by daniel.r...@gmail.com on 25 Apr 2008 at 3:06

• Added labels: ****
• Removed labels: ****

[GoogleCodeExporter]

It would, since we have a test for replication. LGTM anyway.

Original comment by williama...@gmail.com on 25 Apr 2008 at 3:18

• Added labels: ****
• Removed labels: ****

[GoogleCodeExporter]

fixed with revision 1585.

Original comment by jiahotc...@gmail.com on 25 Apr 2008 at 3:29

• Changed state: Fixed
• Added labels: ****
• Removed labels: ****