[Zodb-checkins] CVS: ZODB3/ZEO - stats.py:1.17.6.1.18.4

[Tim Peters]

Update of /cvs-repository/ZODB3/ZEO
In directory cvs.zope.org:/tmp/cvs-serv15109/ZEO

Modified Files:
      Tag: Zope-2_6-branch
	stats.py 
Log Message:
Changed temporal distance to mean the delta across number of load attempts.
This can never be 0 (and isn't <wink>).  I remain unclear on what was
actually being measured before.  Dump changes to:

Histogram of temporal distance
Dist = floor(log2(distance))
Total oids with 1 load: 20,689
Total oids with 2 or more loads: 66,890
 Dist   Count      %  Cumm%
    0   47039   4.32   4.32
    1    3467   0.32   4.63
    2    4139   0.38   5.01
    3    6229   0.57   5.59
    4   11871   1.09   6.67
    5   13536   1.24   7.92
    6   14203   1.30   9.22
    7   11601   1.06  10.28
    8   13623   1.25  11.53
    9   17158   1.57  13.11
   10   34198   3.14  16.25
   11   48331   4.43  20.68
   12   89128   8.18  28.86
   13   95058   8.72  37.58
   14  114634  10.52  48.10
   15  339801  31.18  79.28
   16  140325  12.88  92.16
   17   54998   5.05  97.20
   18   24100   2.21  99.41
   19    6360   0.58 100.00
   20      24   0.00 100.00

Dist=0 there now means two consecutive loads tried to suck up the same
oid.  It's still unclear why that happens so often.


=== ZODB3/ZEO/stats.py 1.17.6.1.18.3 => 1.17.6.1.18.4 ===
--- ZODB3/ZEO/stats.py:1.17.6.1.18.3	Mon Dec  1 21:07:43 2003
+++ ZODB3/ZEO/stats.py	Mon Dec  1 21:39:10 2003
@@ -205,7 +205,7 @@
             if code & 0x70 == 0x20:
                 oids[oid] = oids.get(oid, 0) + 1
                 total_loads += 1
-                refs.setdefault(oid, []).append(datarecords)
+                refs.setdefault(oid, []).append(total_loads)
             if code in (0x00, 0x70):
                 if not quiet:
                     dumpbyinterval(byinterval, h0, he)
@@ -299,8 +299,7 @@
             repeats += 1
             for i in xrange(1, len(L)):
                 distance = L[i] - L[i-1]
-                # XXX How can we get a distance of 0?  We do.
-                lg_distance = int(log(distance or 1) / LOG2)
+                lg_distance = int(log(distance) / LOG2)
                 dist[lg_distance] = dist.get(lg_distance, 0) + 1
 
         L = dist.items()