From: Michael Vrable <mvrable@cs.ucsd.edu>
Date: Thu, 31 May 2007 05:44:11 +0000 (-0700)
Subject: Differentiate between never-before-seen objects and seen-but-expired.
X-Git-Url: http://git.vrable.net/?p=cumulus.git;a=commitdiff_plain;h=cac644d993d130efd8d29539de6557b18c9b737e

Differentiate between never-before-seen objects and seen-but-expired.

We group seen-but-expired objects into different segments, since the fact
that the content has been seen before is an indicator that the data is
long-lived, and grouping by (expected future) age should help increase
segment utilization.
---

diff --git a/localdb.cc b/localdb.cc
index c58549b..4683f4d 100644
--- a/localdb.cc
+++ b/localdb.cc
@@ -221,6 +221,40 @@ ObjectReference LocalDb::FindObject(const string &checksum, int64_t size)
     return ref;
 }
 
+bool LocalDb::IsOldObject(const string &checksum, int64_t size)
+{
+    int rc;
+    sqlite3_stmt *stmt;
+    static const char s[] =
+        "select segmentid, object from block_index "
+        "where checksum = ? and size = ?";
+    const char *tail;
+
+    bool found = false;
+
+    rc = sqlite3_prepare_v2(db, s, strlen(s), &stmt, &tail);
+    if (rc != SQLITE_OK) {
+        return false;
+    }
+
+    sqlite3_bind_text(stmt, 1, checksum.c_str(), checksum.size(),
+                      SQLITE_TRANSIENT);
+    sqlite3_bind_int64(stmt, 2, size);
+
+    rc = sqlite3_step(stmt);
+    if (rc == SQLITE_DONE) {
+        found = false;
+    } else if (rc == SQLITE_ROW) {
+        found = true;
+    } else {
+        fprintf(stderr, "Could not execute SELECT statement!\n");
+    }
+
+    sqlite3_finalize(stmt);
+
+    return found;
+}
+
 void LocalDb::UseObject(const ObjectReference& ref)
 {
     int rc;
diff --git a/localdb.h b/localdb.h
index eb0955d..cebf019 100644
--- a/localdb.h
+++ b/localdb.h
@@ -26,6 +26,7 @@ public:
     void StoreObject(const ObjectReference& ref,
                      const std::string &checksum, int64_t size);
     ObjectReference FindObject(const std::string &checksum, int64_t size);
+    bool IsOldObject(const std::string &checksum, int64_t size);
     void UseObject(const ObjectReference& ref);
 private:
     sqlite3 *db;
diff --git a/scandir.cc b/scandir.cc
index 26b83ed..aa7ce53 100644
--- a/scandir.cc
+++ b/scandir.cc
@@ -145,7 +145,20 @@ int64_t dumpfile(int fd, dictionary &file_info)
         // Store a copy of the object if one does not yet exist
         if (ref.get_segment().size() == 0) {
             LbsObject *o = new LbsObject;
-            o->set_group("data");
+
+            /* We might still have seen this checksum before, if the object was
+             * stored at some time in the past, but we have decided to clean
+             * the segment the object was originally stored in (FindObject will
+             * not return such objects).  When rewriting the object contents,
+             * put it in a separate group, so that old objects get grouped
+             * together.  The hope is that these old objects will continue to
+             * be used in the future, and we obtain segments which will
+             * continue to be well-utilized. */
+            if (db->IsOldObject(block_csum, bytes))
+                o->set_group("compacted");
+            else
+                o->set_group("data");
+
             o->set_data(block_buf, bytes);
             o->write(tss);
             ref = o->get_ref();