diff --git a/indra/llrender/llgl.h b/indra/llrender/llgl.h
--- a/indra/llrender/llgl.h
+++ b/indra/llrender/llgl.h
@@ -40,6 +40,7 @@
 #include "v4math.h"
 #include "llplane.h"
 #include "llgltypes.h"
+#include "llinstancetracker.h"
 
 #include "llglheaders.h"
 #include "glh/glh_linear.h"
@@ -328,9 +329,11 @@
 	Generic pooling scheme for things which use GL names (used for occlusion queries and vertex buffer objects).
 	Prevents thrashing of GL name caches by avoiding calls to glGenFoo and glDeleteFoo.
 */
-class LLGLNamePool
+class LLGLNamePool : public LLInstanceTracker<LLGLNamePool>
 {
 public:
+	typedef LLInstanceTracker<LLGLNamePool> tracker_t;
+
 	struct NameEntry
 	{
 		GLuint name;
@@ -357,13 +360,11 @@
 	GLuint allocate();
 	void release(GLuint name);
 	
-	static void registerPool(LLGLNamePool* pool);
 	static void upkeepPools();
 	static void cleanupPools();
 
 protected:
 	typedef std::vector<LLGLNamePool*> pool_list_t;
-	static pool_list_t sInstances;
 	
 	virtual GLuint allocateName() = 0;
 	virtual void releaseName(GLuint name) = 0;
diff --git a/indra/llrender/llgl.cpp b/indra/llrender/llgl.cpp
--- a/indra/llrender/llgl.cpp
+++ b/indra/llrender/llgl.cpp
@@ -105,7 +105,6 @@
 
 #define LL_GL_NAME_POOLING 0
 
-LLGLNamePool::pool_list_t LLGLNamePool::sInstances;
 std::list<LLGLUpdate*> LLGLUpdate::sGLQ;
 
 #if (LL_WINDOWS || LL_LINUX || LL_SOLARIS)  && !LL_MESA_HEADLESS
@@ -1920,22 +1919,8 @@
 {
 }
 
-void LLGLNamePool::registerPool(LLGLNamePool* pool)
-{
-	pool_list_t::iterator iter = std::find(sInstances.begin(), sInstances.end(), pool);
-	if (iter == sInstances.end())
-	{
-		sInstances.push_back(pool);
-	}
-}
-
 LLGLNamePool::~LLGLNamePool()
 {
-	pool_list_t::iterator iter = std::find(sInstances.begin(), sInstances.end(), this);
-	if (iter != sInstances.end())
-	{
-		sInstances.erase(iter);
-	}
 }
 
 void LLGLNamePool::upkeep()
@@ -2004,20 +1989,22 @@
 void LLGLNamePool::upkeepPools()
 {
 	LLMemType mt(LLMemType::MTYPE_UPKEEP_POOLS);
-	for (pool_list_t::iterator iter = sInstances.begin(); iter != sInstances.end(); ++iter)
+	tracker_t::LLInstanceTrackerScopedGuard guard;
+	for (tracker_t::instance_iter iter = guard.beginInstances(); iter != guard.endInstances(); ++iter)
 	{
-		LLGLNamePool* pool = *iter;
-		pool->upkeep();
+		LLGLNamePool & pool = *iter;
+		pool.upkeep();
 	}
 }
 
 //static
 void LLGLNamePool::cleanupPools()
 {
-	for (pool_list_t::iterator iter = sInstances.begin(); iter != sInstances.end(); ++iter)
+	tracker_t::LLInstanceTrackerScopedGuard guard;
+	for (tracker_t::instance_iter iter = guard.beginInstances(); iter != guard.endInstances(); ++iter)
 	{
-		LLGLNamePool* pool = *iter;
-		pool->cleanup();
+		LLGLNamePool & pool = *iter;
+		pool.cleanup();
 	}
 }
 
diff --git a/indra/llrender/llvertexbuffer.cpp b/indra/llrender/llvertexbuffer.cpp
--- a/indra/llrender/llvertexbuffer.cpp
+++ b/indra/llrender/llvertexbuffer.cpp
@@ -323,10 +323,6 @@
 	}
 
 	sDisableVBOMapping = sEnableVBOs && no_vbo_mapping ;
-	LLGLNamePool::registerPool(&sDynamicVBOPool);
-	LLGLNamePool::registerPool(&sDynamicIBOPool);
-	LLGLNamePool::registerPool(&sStreamVBOPool);
-	LLGLNamePool::registerPool(&sStreamIBOPool);
 }
 
 //static 
diff --git a/indra/newview/llspatialpartition.cpp b/indra/newview/llspatialpartition.cpp
--- a/indra/newview/llspatialpartition.cpp
+++ b/indra/newview/llspatialpartition.cpp
@@ -1635,8 +1635,6 @@
 	mSlopRatio = 0.25f;
 	mInfiniteFarClip = FALSE;
 
-	LLGLNamePool::registerPool(&sQueryPool);
-
 	mOctree = new LLSpatialGroup::OctreeRoot(LLVector3d(0,0,0), 
 											LLVector3d(1,1,1), 
 											NULL);