Revision c47d1771
Added by Andreas Müller about 8 years ago
cdmlib-persistence/src/main/java/eu/etaxonomy/cdm/database/update/CdmUpdater.java | ||
---|---|---|
1 |
// $Id$ |
|
2 |
/** |
|
3 |
* Copyright (C) 2009 EDIT |
|
4 |
* European Distributed Institute of Taxonomy |
|
5 |
* http://www.e-taxonomy.eu |
|
6 |
* |
|
7 |
* The contents of this file are subject to the Mozilla Public License Version 1.1 |
|
8 |
* See LICENSE.TXT at the top of this package for the full license terms. |
|
9 |
*/ |
|
10 |
package eu.etaxonomy.cdm.database.update; |
|
11 |
|
|
12 |
import java.sql.ResultSet; |
|
13 |
|
|
14 |
import org.apache.commons.lang.StringUtils; |
|
15 |
import org.apache.log4j.Logger; |
|
16 |
|
|
17 |
import eu.etaxonomy.cdm.common.monitor.DefaultProgressMonitor; |
|
18 |
import eu.etaxonomy.cdm.common.monitor.IProgressMonitor; |
|
19 |
import eu.etaxonomy.cdm.database.CdmDataSource; |
|
20 |
import eu.etaxonomy.cdm.database.ICdmDataSource; |
|
21 |
import eu.etaxonomy.cdm.database.update.v35_36.SchemaUpdater_35_36; |
|
22 |
import eu.etaxonomy.cdm.database.update.v35_36.TermUpdater_35_36; |
|
23 |
import eu.etaxonomy.cdm.model.metadata.CdmMetaData; |
|
24 |
|
|
25 |
/** |
|
26 |
* This class launches CDM model updates. |
|
27 |
* Currently it splits the update in model updates and defined term related updates by |
|
28 |
* using a {@link ISchemaUpdater schema updater} and a {@link ITermUpdater}. However, this |
|
29 |
* architecture often results in problems and therefore will be replaced by only 1 schema updater. |
|
30 |
* Term updates will be handled differently or also by using the schema updater. |
|
31 |
* <BR> |
|
32 |
* For each new schema version number there usually exists 1 {@link ISchemaUpdater} which is |
|
33 |
* represents a list of schema update steps. {@link ISchemaUpdater schema updaters} are linked |
|
34 |
* to previous updaters which are called, if relevant, previous to the latest updater. |
|
35 |
* So it is possible to upgrade multiple schema version steps in one call. |
|
36 |
* <BR><BR> |
|
37 |
* As said before each {@link ISchemaUpdater schema updater} creates a list of |
|
38 |
* {@link ISchemaUpdaterStep update steps}. |
|
39 |
* <BR><BR> |
|
40 |
* {@link ISchemaUpdater} support progression monitoring with each update step being one task. |
|
41 |
* <BR><BR> |
|
42 |
* ATTENTION: Some steps in the schema update are not transactional by nature. E.g. adding or removing a column |
|
43 |
* to a table in a SQL database can not be handled in a transaction. Therefore failures in |
|
44 |
* certain steps may not lead to a complete rollback of all steps covered by a {@link ISchemaUpdater}. |
|
45 |
* This may lead to a situation where the database becomes inconsistent. |
|
46 |
* <BR><BR> |
|
47 |
* <u>HOW TO ADD A NEW UPDATER?</u><BR> |
|
48 |
* Adding a new updater currently still needs adjustment at multiple places. |
|
49 |
* <BR> |
|
50 |
* <BR>1.) Increment {@link CdmMetaData} schema version number and term version number. |
|
51 |
* <BR>2.) Create a new class instance of {@link SchemaUpdaterBase} (e.g. by copying an old one). |
|
52 |
* <BR>3.) Update startSchemaVersion and endSchemaVersion in this new class, where startSchemaVersion |
|
53 |
* is the old schema version and endSchemaVersion is the new schema version. |
|
54 |
* <BR>4.) Implement {@link ISchemaUpdater#getPreviousUpdater()} and {@link ISchemaUpdater#getNextUpdater()} |
|
55 |
* in a way that the former returns an instance of the previous schema updater and the later returns null (for now). |
|
56 |
* <BR>5.) Go to the previous schema updater class and adjust {@link ISchemaUpdater#getNextUpdater()} |
|
57 |
* in a way that it returns an instance of the newly created updater. |
|
58 |
* <BR>6.) Repeat steps 2.-5. for {@link ITermUpdater} |
|
59 |
* <BR>7.) Adjust {@link #getCurrentSchemaUpdater()} and {@link #getCurrentTermUpdater()} to return |
|
60 |
* instances of the newly created updaters. |
|
61 |
* |
|
62 |
* @see ISchemaUpdater |
|
63 |
* @see ITermUpdater |
|
64 |
* @see ISchemaUpdaterStep |
|
65 |
* @see ITermUpdaterStep |
|
66 |
* |
|
67 |
* @author a.mueller |
|
68 |
* @date 10.09.2010 |
|
69 |
* |
|
70 |
*/ |
|
71 |
public class CdmUpdater { |
|
72 |
private static final Logger logger = Logger.getLogger(CdmUpdater.class); |
|
73 |
|
|
74 |
public static CdmUpdater NewInstance(){ |
|
75 |
return new CdmUpdater(); |
|
76 |
} |
|
77 |
|
|
78 |
|
|
79 |
private ITermUpdater getCurrentTermUpdater() { |
|
80 |
return TermUpdater_35_36.NewInstance(); |
|
81 |
} |
|
82 |
|
|
83 |
/** |
|
84 |
* Returns the current CDM updater |
|
85 |
* @return |
|
86 |
*/ |
|
87 |
private ISchemaUpdater getCurrentSchemaUpdater() { |
|
88 |
return SchemaUpdater_35_36.NewInstance(); |
|
89 |
} |
|
90 |
|
|
91 |
/** |
|
92 |
* @param datasource |
|
93 |
* @param monitor may be <code>null</code> |
|
94 |
* @return |
|
95 |
*/ |
|
96 |
public boolean updateToCurrentVersion(ICdmDataSource datasource, IProgressMonitor monitor){ |
|
97 |
boolean result = true; |
|
98 |
if (monitor == null){ |
|
99 |
monitor = DefaultProgressMonitor.NewInstance(); |
|
100 |
} |
|
101 |
CaseType caseType = CaseType.caseTypeOfDatasource(datasource); |
|
102 |
|
|
103 |
ISchemaUpdater currentSchemaUpdater = getCurrentSchemaUpdater(); |
|
104 |
// TODO do we really always update the terms?? |
|
105 |
ITermUpdater currentTermUpdater = getCurrentTermUpdater(); |
|
106 |
|
|
107 |
int steps = currentSchemaUpdater.countSteps(datasource, monitor, caseType); |
|
108 |
steps += currentTermUpdater.countSteps(datasource, monitor, caseType); |
|
109 |
steps++; //for hibernate_sequences update |
|
110 |
|
|
111 |
String taskName = "Update to schema version " + currentSchemaUpdater.getTargetVersion() + " and to term version " + currentTermUpdater.getTargetVersion(); //+ currentSchemaUpdater.getVersion(); |
|
112 |
monitor.beginTask(taskName, steps); |
|
113 |
|
|
114 |
try { |
|
115 |
datasource.startTransaction(); |
|
116 |
result &= currentSchemaUpdater.invoke(datasource, monitor, caseType); |
|
117 |
if (result == true){ |
|
118 |
result &= currentTermUpdater.invoke(datasource, monitor, caseType); |
|
119 |
updateHibernateSequence(datasource, monitor, caseType); |
|
120 |
} |
|
121 |
if (result == false){ |
|
122 |
datasource.rollback(); |
|
123 |
}else{ |
|
124 |
datasource.commitTransaction(); |
|
125 |
} |
|
126 |
|
|
127 |
} catch (Exception e) { |
|
128 |
result = false; |
|
129 |
monitor.warning("Stopped schema updater"); |
|
130 |
} finally { |
|
131 |
String message = "Update finished " + (result ? "successfully" : "with ERRORS"); |
|
132 |
monitor.subTask(message); |
|
133 |
if (!result){ |
|
134 |
monitor.warning(message); |
|
135 |
monitor.setCanceled(true); |
|
136 |
}else{ |
|
137 |
monitor.done(); |
|
138 |
} |
|
139 |
logger.info(message); |
|
140 |
} |
|
141 |
|
|
142 |
return result; |
|
143 |
} |
|
144 |
|
|
145 |
|
|
146 |
|
|
147 |
/** |
|
148 |
* Updating terms often inserts new terms, vocabularies and representations. |
|
149 |
* Therefore the counter in hibernate_sequences must be increased. |
|
150 |
* We do this once at the end of term updating. |
|
151 |
* @param caseType |
|
152 |
* @return true if update was successful, false otherwise |
|
153 |
*/ |
|
154 |
private boolean updateHibernateSequence(ICdmDataSource datasource, IProgressMonitor monitor, CaseType caseType) { |
|
155 |
boolean result = true; |
|
156 |
monitor.subTask("Update hibernate sequences"); |
|
157 |
try { |
|
158 |
String sql = "SELECT * FROM hibernate_sequences "; |
|
159 |
ResultSet rs = datasource.executeQuery(sql); |
|
160 |
while (rs.next()){ |
|
161 |
String table = rs.getString("sequence_name"); |
|
162 |
Integer val = rs.getInt("next_val"); |
|
163 |
result &= updateSingleValue(datasource,monitor, table, val, caseType); |
|
164 |
} |
|
165 |
} catch (Exception e) { |
|
166 |
String message = "Exception occurred when trying to update hibernate_sequences table: " + e.getMessage(); |
|
167 |
monitor.warning(message, e); |
|
168 |
logger.error(message); |
|
169 |
result = false; |
|
170 |
}finally{ |
|
171 |
monitor.worked(1); |
|
172 |
} |
|
173 |
return result; |
|
174 |
} |
|
175 |
|
|
176 |
/** |
|
177 |
* |
|
178 |
* @param datasource |
|
179 |
* @param monitor |
|
180 |
* @param table |
|
181 |
* @param oldVal |
|
182 |
* @param caseType |
|
183 |
* @return |
|
184 |
*/ |
|
185 |
private boolean updateSingleValue(ICdmDataSource datasource, IProgressMonitor monitor, String table, Integer oldVal, CaseType caseType){ |
|
186 |
if (table.equals("default")){ //found in flora central africa test database |
|
187 |
return true; |
|
188 |
} |
|
189 |
try { |
|
190 |
Integer newVal; |
|
191 |
try { |
|
192 |
String sql = " SELECT max(id) FROM %s "; |
|
193 |
newVal = (Integer)datasource.getSingleValue(String.format(sql, caseType.transformTo(table))); |
|
194 |
} catch (Exception e) { |
|
195 |
String message = "Could not retrieve max value for table '%s'. Will not update hibernate_sequence for this table. " + |
|
196 |
"Usually this will not cause problems, however, if new data has been added to " + |
|
197 |
"this table by the update script one may encounter 'unique identifier' " + |
|
198 |
"exceptions when trying to add further data."; |
|
199 |
monitor.warning(String.format(message,table), e); |
|
200 |
//TODO |
|
201 |
return true; |
|
202 |
} |
|
203 |
|
|
204 |
if (newVal != null){ |
|
205 |
//This is how {@link PooledOptimizer#generate(org.hibernate.id.enhanced.AccessCallback)} works |
|
206 |
//it substracts the increment size from the value in hibernate_sequences to get the initial value. |
|
207 |
//Haven't checked why. |
|
208 |
//For the correct increment size see eu.etaxonomy.cdm.model.common.package-info.java |
|
209 |
int incrementSize = 10; |
|
210 |
newVal = newVal + incrementSize; |
|
211 |
if (newVal != null && newVal >= oldVal){ |
|
212 |
String sql = " UPDATE hibernate_sequences " + |
|
213 |
" SET next_val = %d " + |
|
214 |
" WHERE sequence_name = '%s' "; |
|
215 |
datasource.executeUpdate(String.format(sql, newVal + 1 , table) ); |
|
216 |
} |
|
217 |
} |
|
218 |
return true; |
|
219 |
} catch (Exception e) { |
|
220 |
String message = "Exception occurred when trying to read or update hibernate_sequences table for value " + table + ": " + e.getMessage(); |
|
221 |
monitor.warning(message, e); |
|
222 |
logger.error(message); |
|
223 |
return false; |
|
224 |
} |
|
225 |
|
|
226 |
} |
|
227 |
|
|
228 |
|
|
229 |
|
|
230 |
/** |
|
231 |
* |
|
232 |
* |
|
233 |
* @param args SERVER DB_NAME1[,DB_NAME2,...] [USER] [PASSWORD] [PORT] |
|
234 |
*/ |
|
235 |
public static void main(String[] args) { |
|
236 |
// logger.warn("main method not yet fully implemented (only works with mysql!!!)"); |
|
237 |
// if(args.length < 2){ |
|
238 |
// logger.error("Arguments missing: server database [username [password]]"); |
|
239 |
// } |
|
240 |
//TODO better implementation |
|
241 |
CdmUpdater myUpdater = new CdmUpdater(); |
|
242 |
System.out.println("CdmUpdater\nArguments: SERVER DB_NAME1[,DB_NAME2,...] [USER] [PASSWORD] [PORT]"); |
|
243 |
String server = args[0]; |
|
244 |
String database = args[1]; |
|
245 |
String[] databaseNames = StringUtils.split(database, ','); |
|
246 |
String username = args.length > 2 ? args[2] : null; |
|
247 |
String password = args.length > 3 ? args[3] : null; |
|
248 |
int port = 3306; |
|
249 |
if( args.length > 4){ |
|
250 |
try { |
|
251 |
port = Integer.parseInt(args[4]); |
|
252 |
} catch (Exception e) { |
|
253 |
// ignore |
|
254 |
} |
|
255 |
} |
|
256 |
System.out.println("Number of databases to update: " + databaseNames.length); |
|
257 |
for(String dnName : databaseNames){ |
|
258 |
System.out.println(dnName + " UPDATE ..."); |
|
259 |
ICdmDataSource dataSource = CdmDataSource.NewMySqlInstance(server, dnName, port, username, password, null); |
|
260 |
boolean success = myUpdater.updateToCurrentVersion(dataSource, null); |
|
261 |
System.out.println(dnName + " DONE " + (success ? "successfully" : "with ERRORS")); |
|
262 |
System.out.println("===================================================================="); |
|
263 |
|
|
264 |
} |
|
265 |
} |
|
266 |
|
|
267 |
} |
|
1 |
// $Id$ |
|
2 |
/** |
|
3 |
* Copyright (C) 2009 EDIT |
|
4 |
* European Distributed Institute of Taxonomy |
|
5 |
* http://www.e-taxonomy.eu |
|
6 |
* |
|
7 |
* The contents of this file are subject to the Mozilla Public License Version 1.1 |
|
8 |
* See LICENSE.TXT at the top of this package for the full license terms. |
|
9 |
*/ |
|
10 |
package eu.etaxonomy.cdm.database.update; |
|
11 |
|
|
12 |
import java.sql.ResultSet; |
|
13 |
|
|
14 |
import org.apache.commons.lang.StringUtils; |
|
15 |
import org.apache.log4j.Logger; |
|
16 |
|
|
17 |
import eu.etaxonomy.cdm.common.monitor.DefaultProgressMonitor; |
|
18 |
import eu.etaxonomy.cdm.common.monitor.IProgressMonitor; |
|
19 |
import eu.etaxonomy.cdm.database.CdmDataSource; |
|
20 |
import eu.etaxonomy.cdm.database.ICdmDataSource; |
|
21 |
import eu.etaxonomy.cdm.database.update.v35_36.SchemaUpdater_35_36; |
|
22 |
import eu.etaxonomy.cdm.database.update.v35_36.TermUpdater_35_36; |
|
23 |
import eu.etaxonomy.cdm.model.metadata.CdmMetaData; |
|
24 |
|
|
25 |
/** |
|
26 |
* This class launches CDM model updates. |
|
27 |
* Currently it splits the update in model updates and defined term related updates by |
|
28 |
* using a {@link ISchemaUpdater schema updater} and a {@link ITermUpdater}. However, this |
|
29 |
* architecture often results in problems and therefore will be replaced by only 1 schema updater. |
|
30 |
* Term updates will be handled differently or also by using the schema updater. |
|
31 |
* <BR> |
|
32 |
* For each new schema version number there usually exists 1 {@link ISchemaUpdater} which is |
|
33 |
* represents a list of schema update steps. {@link ISchemaUpdater schema updaters} are linked |
|
34 |
* to previous updaters which are called, if relevant, previous to the latest updater. |
|
35 |
* So it is possible to upgrade multiple schema version steps in one call. |
|
36 |
* <BR><BR> |
|
37 |
* As said before each {@link ISchemaUpdater schema updater} creates a list of |
|
38 |
* {@link ISchemaUpdaterStep update steps}. |
|
39 |
* <BR><BR> |
|
40 |
* {@link ISchemaUpdater} support progression monitoring with each update step being one task. |
|
41 |
* <BR><BR> |
|
42 |
* ATTENTION: Some steps in the schema update are not transactional by nature. E.g. adding or removing a column |
|
43 |
* to a table in a SQL database can not be handled in a transaction. Therefore failures in |
|
44 |
* certain steps may not lead to a complete rollback of all steps covered by a {@link ISchemaUpdater}. |
|
45 |
* This may lead to a situation where the database becomes inconsistent. |
|
46 |
* <BR><BR> |
|
47 |
* <u>HOW TO ADD A NEW UPDATER?</u><BR> |
|
48 |
* Adding a new updater currently still needs adjustment at multiple places. |
|
49 |
* <BR> |
|
50 |
* <BR>1.) Increment {@link CdmMetaData} schema version number and term version number. |
|
51 |
* <BR>2.) Create a new class instance of {@link SchemaUpdaterBase} (e.g. by copying an old one). |
|
52 |
* <BR>3.) Update startSchemaVersion and endSchemaVersion in this new class, where startSchemaVersion |
|
53 |
* is the old schema version and endSchemaVersion is the new schema version. |
|
54 |
* <BR>4.) Implement {@link ISchemaUpdater#getPreviousUpdater()} and {@link ISchemaUpdater#getNextUpdater()} |
|
55 |
* in a way that the former returns an instance of the previous schema updater and the later returns null (for now). |
|
56 |
* <BR>5.) Go to the previous schema updater class and adjust {@link ISchemaUpdater#getNextUpdater()} |
|
57 |
* in a way that it returns an instance of the newly created updater. |
|
58 |
* <BR>6.) Repeat steps 2.-5. for {@link ITermUpdater} |
|
59 |
* <BR>7.) Adjust {@link #getCurrentSchemaUpdater()} and {@link #getCurrentTermUpdater()} to return |
|
60 |
* instances of the newly created updaters. |
|
61 |
* |
|
62 |
* @see ISchemaUpdater |
|
63 |
* @see ITermUpdater |
|
64 |
* @see ISchemaUpdaterStep |
|
65 |
* @see ITermUpdaterStep |
|
66 |
* |
|
67 |
* @author a.mueller |
|
68 |
* @date 10.09.2010 |
|
69 |
* |
|
70 |
*/ |
|
71 |
public class CdmUpdater { |
|
72 |
private static final Logger logger = Logger.getLogger(CdmUpdater.class); |
|
73 |
|
|
74 |
public static CdmUpdater NewInstance(){ |
|
75 |
return new CdmUpdater(); |
|
76 |
} |
|
77 |
|
|
78 |
|
|
79 |
private ITermUpdater getCurrentTermUpdater() { |
|
80 |
return TermUpdater_35_36.NewInstance(); |
|
81 |
} |
|
82 |
|
|
83 |
/** |
|
84 |
* Returns the current CDM updater |
|
85 |
* @return |
|
86 |
*/ |
|
87 |
private ISchemaUpdater getCurrentSchemaUpdater() { |
|
88 |
return SchemaUpdater_35_36.NewInstance(); |
|
89 |
} |
|
90 |
|
|
91 |
/** |
|
92 |
* @param datasource |
|
93 |
* @param monitor may be <code>null</code> |
|
94 |
* @return |
|
95 |
*/ |
|
96 |
public boolean updateToCurrentVersion(ICdmDataSource datasource, IProgressMonitor monitor){ |
|
97 |
boolean result = true; |
|
98 |
if (monitor == null){ |
|
99 |
monitor = DefaultProgressMonitor.NewInstance(); |
|
100 |
} |
|
101 |
CaseType caseType = CaseType.caseTypeOfDatasource(datasource); |
|
102 |
|
|
103 |
ISchemaUpdater currentSchemaUpdater = getCurrentSchemaUpdater(); |
|
104 |
// TODO do we really always update the terms?? |
|
105 |
ITermUpdater currentTermUpdater = getCurrentTermUpdater(); |
|
106 |
|
|
107 |
int steps = currentSchemaUpdater.countSteps(datasource, monitor, caseType); |
|
108 |
steps += currentTermUpdater.countSteps(datasource, monitor, caseType); |
|
109 |
steps++; //for hibernate_sequences update |
|
110 |
|
|
111 |
String taskName = "Update to schema version " + currentSchemaUpdater.getTargetVersion() + " and to term version " + currentTermUpdater.getTargetVersion(); //+ currentSchemaUpdater.getVersion(); |
|
112 |
monitor.beginTask(taskName, steps); |
|
113 |
|
|
114 |
try { |
|
115 |
datasource.startTransaction(); |
|
116 |
result &= currentSchemaUpdater.invoke(datasource, monitor, caseType); |
|
117 |
if (result == true){ |
|
118 |
result &= currentTermUpdater.invoke(datasource, monitor, caseType); |
|
119 |
updateHibernateSequence(datasource, monitor, caseType); |
|
120 |
} |
|
121 |
if (result == false){ |
|
122 |
datasource.rollback(); |
|
123 |
}else{ |
|
124 |
datasource.commitTransaction(); |
|
125 |
} |
|
126 |
|
|
127 |
} catch (Exception e) { |
|
128 |
result = false; |
|
129 |
monitor.warning("Stopped schema updater"); |
|
130 |
} finally { |
|
131 |
String message = "Update finished " + (result ? "successfully" : "with ERRORS"); |
|
132 |
monitor.subTask(message); |
|
133 |
if (!result){ |
|
134 |
monitor.warning(message); |
|
135 |
monitor.setCanceled(true); |
|
136 |
}else{ |
|
137 |
monitor.done(); |
|
138 |
} |
|
139 |
logger.info(message); |
|
140 |
} |
|
141 |
|
|
142 |
return result; |
|
143 |
} |
|
144 |
|
|
145 |
|
|
146 |
|
|
147 |
/** |
|
148 |
* Updating terms often inserts new terms, vocabularies and representations. |
|
149 |
* Therefore the counter in hibernate_sequences must be increased. |
|
150 |
* We do this once at the end of term updating. |
|
151 |
* @param caseType |
|
152 |
* @return true if update was successful, false otherwise |
|
153 |
*/ |
|
154 |
private boolean updateHibernateSequence(ICdmDataSource datasource, IProgressMonitor monitor, CaseType caseType) { |
|
155 |
boolean result = true; |
|
156 |
monitor.subTask("Update hibernate sequences"); |
|
157 |
try { |
|
158 |
String sql = "SELECT * FROM hibernate_sequences "; |
|
159 |
ResultSet rs = datasource.executeQuery(sql); |
|
160 |
while (rs.next()){ |
|
161 |
String table = rs.getString("sequence_name"); |
|
162 |
Integer val = rs.getInt("next_val"); |
|
163 |
result &= updateSingleValue(datasource, monitor, table, val, caseType); |
|
164 |
} |
|
165 |
} catch (Exception e) { |
|
166 |
String message = "Exception occurred when trying to update hibernate_sequences table: " + e.getMessage(); |
|
167 |
monitor.warning(message, e); |
|
168 |
logger.error(message); |
|
169 |
result = false; |
|
170 |
}finally{ |
|
171 |
monitor.worked(1); |
|
172 |
} |
|
173 |
return result; |
|
174 |
} |
|
175 |
|
|
176 |
/** |
|
177 |
* |
|
178 |
* @param datasource |
|
179 |
* @param monitor |
|
180 |
* @param table |
|
181 |
* @param oldVal |
|
182 |
* @param caseType |
|
183 |
* @return |
|
184 |
*/ |
|
185 |
private boolean updateSingleValue(ICdmDataSource datasource, IProgressMonitor monitor, String table, Integer oldVal, CaseType caseType){ |
|
186 |
if (table.equals("default")){ //found in flora central africa test database |
|
187 |
return true; |
|
188 |
} |
|
189 |
try { |
|
190 |
Integer newVal; |
|
191 |
try { |
|
192 |
String id = table.equalsIgnoreCase("AuditEvent")? "revisionNumber" : "id"; |
|
193 |
String sql = " SELECT max(%s) FROM %s "; |
|
194 |
newVal = (Integer)datasource.getSingleValue(String.format(sql, id, caseType.transformTo(table))); |
|
195 |
} catch (Exception e) { |
|
196 |
String message = "Could not retrieve max value for table '%s'. Will not update hibernate_sequence for this table. " + |
|
197 |
"Usually this will not cause problems, however, if new data has been added to " + |
|
198 |
"this table by the update script one may encounter 'unique identifier' " + |
|
199 |
"exceptions when trying to add further data."; |
|
200 |
monitor.warning(String.format(message,table), e); |
|
201 |
//TODO |
|
202 |
return true; |
|
203 |
} |
|
204 |
|
|
205 |
if (newVal != null){ |
|
206 |
//This is how {@link PooledOptimizer#generate(org.hibernate.id.enhanced.AccessCallback)} works |
|
207 |
//it substracts the increment size from the value in hibernate_sequences to get the initial value. |
|
208 |
//Haven't checked why. |
|
209 |
//For the correct increment size see eu.etaxonomy.cdm.model.common.package-info.java |
|
210 |
int incrementSize = 10; |
|
211 |
newVal = newVal + incrementSize; |
|
212 |
if (newVal != null && newVal >= oldVal){ |
|
213 |
String sql = " UPDATE hibernate_sequences " + |
|
214 |
" SET next_val = %d " + |
|
215 |
" WHERE sequence_name = '%s' "; |
|
216 |
datasource.executeUpdate(String.format(sql, newVal + 1 , table) ); |
|
217 |
} |
|
218 |
} |
|
219 |
return true; |
|
220 |
} catch (Exception e) { |
|
221 |
String message = "Exception occurred when trying to read or update hibernate_sequences table for value " + table + ": " + e.getMessage(); |
|
222 |
monitor.warning(message, e); |
|
223 |
logger.error(message); |
|
224 |
return false; |
|
225 |
} |
|
226 |
|
|
227 |
} |
|
228 |
|
|
229 |
|
|
230 |
|
|
231 |
/** |
|
232 |
* |
|
233 |
* |
|
234 |
* @param args SERVER DB_NAME1[,DB_NAME2,...] [USER] [PASSWORD] [PORT] |
|
235 |
*/ |
|
236 |
public static void main(String[] args) { |
|
237 |
// logger.warn("main method not yet fully implemented (only works with mysql!!!)"); |
|
238 |
// if(args.length < 2){ |
|
239 |
// logger.error("Arguments missing: server database [username [password]]"); |
|
240 |
// } |
|
241 |
//TODO better implementation |
|
242 |
CdmUpdater myUpdater = new CdmUpdater(); |
|
243 |
System.out.println("CdmUpdater\nArguments: SERVER DB_NAME1[,DB_NAME2,...] [USER] [PASSWORD] [PORT]"); |
|
244 |
String server = args[0]; |
|
245 |
String database = args[1]; |
|
246 |
String[] databaseNames = StringUtils.split(database, ','); |
|
247 |
String username = args.length > 2 ? args[2] : null; |
|
248 |
String password = args.length > 3 ? args[3] : null; |
|
249 |
int port = 3306; |
|
250 |
if( args.length > 4){ |
|
251 |
try { |
|
252 |
port = Integer.parseInt(args[4]); |
|
253 |
} catch (Exception e) { |
|
254 |
// ignore |
|
255 |
} |
|
256 |
} |
|
257 |
System.out.println("Number of databases to update: " + databaseNames.length); |
|
258 |
for(String dnName : databaseNames){ |
|
259 |
System.out.println(dnName + " UPDATE ..."); |
|
260 |
ICdmDataSource dataSource = CdmDataSource.NewMySqlInstance(server, dnName, port, username, password, null); |
|
261 |
boolean success = myUpdater.updateToCurrentVersion(dataSource, null); |
|
262 |
System.out.println(dnName + " DONE " + (success ? "successfully" : "with ERRORS")); |
|
263 |
System.out.println("===================================================================="); |
|
264 |
|
|
265 |
} |
|
266 |
} |
|
267 |
|
|
268 |
} |
Also available in: Unified diff
Handle AuditEvent hibernate sequence update in CdmUpdater correctly
#4716