1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
|
# See the file LICENSE for redistribution information.
#
# Copyright (c) 2009, 2013 Oracle and/or its affiliates. All rights reserved.
#
# TEST rep082
# TEST Sending replication requests to correct master site.
# TEST
# TEST Regression test for a bug [#16592] where a client could send an
# TEST UPDATE_REQ to another client instead of the master.
#
proc rep082 { method { niter 200 } { tnum "082" } args } {
source ./include.tcl
global repfiles_in_memory
if { $checking_valid_methods } {
return "ALL"
}
set args [convert_args $method $args]
set msg2 "and on-disk replication files"
if { $repfiles_in_memory } {
set msg2 "and in-memory replication files"
}
puts "Rep$tnum: ($method) Test that\
client doesn't send UPDATE_REQ to another client $msg2."
rep082_sub $method $niter $tnum $args
}
proc rep082_sub { method niter tnum largs } {
global testdir
global util_path
global repfiles_in_memory
global rep_verbose
global verbose_type
set verbargs ""
if { $rep_verbose == 1 } {
set verbargs " -verbose {$verbose_type on} "
}
set repmemargs ""
if { $repfiles_in_memory } {
set repmemargs "-rep_inmem_files "
}
env_cleanup $testdir
replsetup $testdir/MSGQUEUEDIR
file mkdir [set dirA $testdir/A]
file mkdir [set dirB $testdir/B]
file mkdir [set dirC $testdir/C]
set pagesize 4096
append largs " -pagesize $pagesize "
set log_buf [expr $pagesize * 2]
set log_max [expr $log_buf * 4]
repladd 1
set env_A_cmd "berkdb_env_noerr -create -txn $verbargs $repmemargs \
-log_buffer $log_buf -log_max $log_max -errpfx SITE_A \
-home $dirA -rep_transport \[list 1 replsend\]"
set envs(A) [eval $env_A_cmd -rep_master]
# Open a client
repladd 2
set env_B_cmd "berkdb_env_noerr -create -txn $verbargs $repmemargs \
-log_buffer $log_buf -log_max $log_max -errpfx SITE_B \
-home $dirB -rep_transport \[list 2 replsend\]"
set envs(B) [eval $env_B_cmd -rep_client]
# Open 2nd client
repladd 3
set env_C_cmd "berkdb_env_noerr -create -txn $verbargs $repmemargs \
-log_buffer $log_buf -log_max $log_max -errpfx SITE_C \
-home $dirC -rep_transport \[list 3 replsend\]"
set envs(C) [eval $env_C_cmd -rep_client]
# Bring the clients online by processing the startup messages.
set envlist "{$envs(A) 1} {$envs(B) 2} {$envs(C) 3}"
process_msgs $envlist
# Run rep_test in the master (and update clients).
puts "\tRep$tnum.a: populate initial portion of log."
eval rep_test $method $envs(A) NULL $niter 0 0 0 $largs
process_msgs $envlist
$envs(A) close
$envs(B) close
$envs(C) close
# At this point, we have a first section of the log history that was
# produced at master site A, and is replicated to both other sites. Now
# let's produce a second section of history, also produced at master
# site A, but only replicated to site C; make sure this second section
# spans a log file boundary. Archive log files at site C, so that we
# make sure that site C has only a fraction of this second section.
#
set envs(A) [eval $env_A_cmd -recover -rep_master]
set envs(C) [eval $env_C_cmd -recover -rep_client]
replclear 2
process_msgs "{$envs(A) 1} {$envs(C) 3}"
set last_client_log [get_logfile $envs(C) last]
set stop 0
set start 0
set count 0
while { $stop == 0 } {
incr count
# Run rep_test in the master (don't update client).
puts "\tRep$tnum.b: Fill log until next log file."
incr start $niter
eval rep_test $method $envs(A) NULL $niter $start $start 0 $largs
replclear 2
process_msgs "{$envs(A) 1} {$envs(C) 3}"
puts "\tRep$tnum.c: Run db_archive on client C."
exec $util_path/db_archive -d -h $dirC
set new_last_log [get_logfile $envs(C) last]
if { $new_last_log > $last_client_log } {
set stop 1
}
}
# Now make site B become the master. Since site B was not running
# during the last phase, it does not have any of the "second section of
# log history" that we produced in that phase. So site A will have to
# throw away those transactions in order to sync up with B. HOWEVER,
# site B will now generate yet another new section of log history, which
# is identical to the set of transactions generated a moment ago at site
# A. In other words, although this is the third section of history to
# be generated, we have arranged to have it completely replace the
# second section, and to have it exactly match! Note that we leave site
# C out of the picture during this phase.
#
$envs(A) close
$envs(C) close
set envs(B) [eval $env_B_cmd -recover -rep_master]
set envs(A) [eval $env_A_cmd -recover -rep_client]
set start 0
while {$count > 0} {
puts "\tRep$tnum.d: Running rep_test in replicated env."
incr start $niter
eval rep_test $method $envs(B) NULL $niter $start $start 0 $largs
replclear 3
process_msgs "{$envs(A) 1} {$envs(B) 2}"
incr count -1
}
# Now start up site C again, but configure it to rely on site A for
# client-to-client synchronization. Recall the known contents of site
# C's transaction log: it has a partial copy of the "second section" of
# log history (it has the end of that section, but not the beginning).
# The transactions in this log will have the same LSN's as are currently
# in place at sites A and B (which, remember, were produced by the
# identical "third section" of history), but the commit record contents
# won't exactly match, because the third section was produced by master
# site B.
#
# During the verify dance, client C will continue to walk back the log,
# finding commit records which find matching LSNs at A/B, but no
# matching contents. When it hits the archived log file boundary it
# will have to give up without having found a match. Thus we have
# produced a situation where an incoming VERIFY message from another
# client (site A) results in client C sending an UPDATE_REQ. We want to
# make sure that client C sends the UPDATE_REQ to the master, rather
# than blindly sending to the same site that produced the VERIFY
# message.
#
puts "\tRep$tnum.e: start client C, with A as peer."
set env_C_cmd "berkdb_env_noerr -create -txn $verbargs \
-log_buffer $log_buf -log_max $log_max -errpfx SITE_C \
-home $dirC -rep_transport \[list 3 rep082_send\]"
set envs(C) [eval $env_C_cmd -recover -rep_client]
process_msgs "{$envs(A) 1} {$envs(B) 2} {$envs(C) 3}"
$envs(C) close
$envs(A) close
$envs(B) close
replclose $testdir/MSGQUEUEDIR
}
# We use this special-purpose wrapper send function only in the very last phase
# of the test, and only at site C. Before that we just use the normal send
# function as usual. Since we know exactly what sites are in what roles, we can
# simply hard-code the EID numbers: site B (2) is the master, and site A (1) is
# the desired target site for c2c synchronization.
#
proc rep082_send { control rec fromid toid flags lsn } {
if {$toid == 2 && [lsearch $flags "rerequest"] == -1 \
&& [lsearch $flags "any"] != -1} {
set toid 1
}
replsend $control $rec $fromid $toid $flags $lsn
}
|
