Skip to contentSkip to conversation

Fix CLUSTERED flag in new _vg_read_for_update() and vg_read().

PATCH New Dave Wysochanski 2 participants 2 messages May 26, 2009

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 Fix CLUSTERED flag in new _vg_read_for_update() and vg_read().

In the new _vg_read_for_update(), we always do the check for CLUSTERED vg status flag after reading the volume group. Thus, no need to set the flag in vg_read() or clear it later before calling _vg_bad_status_bits().

Also, add back in the !lockingfailed() as part of the CLUSTERED flag check. It's unclear why it was removed when the check was moved from _vg_bad_status_bits() to inside _vg_lock_and_read().

---
 lib/metadata/metadata.c |    8 ++++----
 1 files changed, 4 insertions(+), 4 deletions(-)
lib/metadata/metadata.c +4 -4
index f16cd55..55c4578 100644
@@ -2763,7 +2763,8 @@ static vg_t *_vg_lock_and_read(struct cmd_context *cmd, const char *vg_name,
 		goto_bad;
 	}
 
-	if (vg_is_clustered(vg) && !locking_is_clustered()) {
+	if ((vg_is_clustered(vg)) && !locking_is_clustered() &&
+	    !lockingfailed()) {
 		log_error("Skipping clustered volume group %s", vg->name);
 		failure |= FAILED_CLUSTERED;
 		goto_bad;
@@ -2779,9 +2780,8 @@ static vg_t *_vg_lock_and_read(struct cmd_context *cmd, const char *vg_name,
 			goto_bad;
 		}
 	}
-	
 
-	failure |= _vg_bad_status_bits(vg, status_flags & ~CLUSTERED);
+	failure |= _vg_bad_status_bits(vg, status_flags);
 	if (failure)
 		goto_bad;
 
@@ -2828,7 +2828,7 @@ bad:
 vg_t *vg_read(struct cmd_context *cmd, const char *vg_name,
 	      const char *vgid, uint32_t flags)
 {
-	uint32_t status = CLUSTERED;
+	uint32_t status = 0;
 	uint32_t lock_flags = LCK_VG_READ;
 
 	if (flags & READ_FOR_UPDATE) {
Signature · 1 line
1.6.0.6
#1 lore raw

PATCH 1/1 Fix CLUSTERED flag in new _vg_read_for_update() and vg_read().

Quoted 3 lines · Dave Wysochanski

In the new _vg_read_for_update(), we always do the check for CLUSTERED vg status flag after reading the volume group. Thus, no need to set the flag in vg_read() or clear it later before calling _vg_bad_status_bits().

Ok, makes sense.

Quoted 3 lines

Also, add back in the !lockingfailed() as part of the CLUSTERED flag check. It's unclear why it was removed when the check was moved from _vg_bad_status_bits() to inside _vg_lock_and_read().

There is a patch floating somewhere removing lockingfailed() altogether, since it's an ugly piece of global state (a rampant modularity violation, if you will). But I concede to have it put back, although I don't really know what its effect here would be, anyway (don't report errors when cluster locking failed? why?).

Yours,
Petr.

Signature · 6 lines
Peter Rockai | me()mornfall!net | prockai()redhat!com
 http://blog.mornfall.net | http://web.mornfall.net

"In My Egotistical Opinion, most people's C programs should be
 indented six feet downward and covered with dirt."
     -- Blair P. Houghton on the subject of C program indentation