[CASSANDRA-15794] Upgraded C* (4.x) fail to start because of Compact Tables & dropping compact tables in downgraded C* (3.11.4) introduces non-existent columns

• Type: Bug
• Status: Resolved
• Priority: Normal
• Resolution: Information Provided
• Severity: Normal
• Discovered By: User Report
• Platform: All
• Since Version: 3.11.4
• Assignee: Unassigned
• Reporter: Zhuqi Jin
• Created: 07/May/20 01:54
• Updated: 25/Jun/20 14:53
• Resolved: 18/May/20 09:58

Description

We tried to test upgrading a 3.11.4 C* cluster to 4.x and run into the following problems. 

• We started a single 3.11.4 C* node. 
• We ran cassandra-stress like this

./cassandra-stress write n = 30 -rate threads = 10 -node  172.17.0.2 

• We stopped this node, and started a C* node running C* compiled from trunk (git commit: e394dc0bb32f612a476269010930c617dd1ed3cb)
• New C* failed to start with the following error message

ERROR [main] 2020-05-07 00:58:18,503 CassandraDaemon.java:245 - Error while loading schema: ERROR [main] 2020-05-07 00:58:18,503 CassandraDaemon.java:245 - Error while loading schema: java.lang.IllegalArgumentException: Compact Tables are not allowed in Cassandra starting with 4.0 version. Use `ALTER ... DROP COMPACT STORAGE` command supplied in 3.x/3.11 Cassandra in order to migrate off Compact Storage. at org.apache.cassandra.schema.SchemaKeyspace.fetchTable(SchemaKeyspace.java:965) at org.apache.cassandra.schema.SchemaKeyspace.fetchTables(SchemaKeyspace.java:924) at org.apache.cassandra.schema.SchemaKeyspace.fetchKeyspace(SchemaKeyspace.java:883) at org.apache.cassandra.schema.SchemaKeyspace.fetchKeyspacesWithout(SchemaKeyspace.java:874) at org.apache.cassandra.schema.SchemaKeyspace.fetchNonSystemKeyspaces(SchemaKeyspace.java:862) at org.apache.cassandra.schema.Schema.loadFromDisk(Schema.java:102) at org.apache.cassandra.schema.Schema.loadFromDisk(Schema.java:91) at org.apache.cassandra.service.CassandraDaemon.setup(CassandraDaemon.java:241) at org.apache.cassandra.service.CassandraDaemon.activate(CassandraDaemon.java:653) at org.apache.cassandra.service.CassandraDaemon.main(CassandraDaemon.java:770)Exception (java.lang.IllegalArgumentException) encountered during startup: Compact Tables are not allowed in Cassandra starting with 4.0 version. Use `ALTER ... DROP COMPACT STORAGE` command supplied in 3.x/3.11 Cassandra in order to migrate off Compact Storage.ERROR [main] 2020-05-07 00:58:18,520 CassandraDaemon.java:792 - Exception encountered during startupjava.lang.IllegalArgumentException: Compact Tables are not allowed in Cassandra starting with 4.0 version. Use `ALTER ... DROP COMPACT STORAGE` command supplied in 3.x/3.11 Cassandra in order to migrate off Compact Storage. at org.apache.cassandra.schema.SchemaKeyspace.fetchTable(SchemaKeyspace.java:965) at org.apache.cassandra.schema.SchemaKeyspace.fetchTables(SchemaKeyspace.java:924) at org.apache.cassandra.schema.SchemaKeyspace.fetchKeyspace(SchemaKeyspace.java:883) at org.apache.cassandra.schema.SchemaKeyspace.fetchKeyspacesWithout(SchemaKeyspace.java:874) at org.apache.cassandra.schema.SchemaKeyspace.fetchNonSystemKeyspaces(SchemaKeyspace.java:862) at org.apache.cassandra.schema.Schema.loadFromDisk(Schema.java:102) at org.apache.cassandra.schema.Schema.loadFromDisk(Schema.java:91) at org.apache.cassandra.service.CassandraDaemon.setup(CassandraDaemon.java:241) at org.apache.cassandra.service.CassandraDaemon.activate(CassandraDaemon.java:653) at org.apache.cassandra.service.CassandraDaemon.main(CassandraDaemon.java:770)

• We stopped the trunk version C* and started the 3.11.4 version C*. 
• 3.11.4 C* failed to start with the following error messages: 

Exception (java.lang.IllegalStateException) encountered during startup: Unknown commitlog version 7Exception (java.lang.IllegalStateException) encountered during startup: Unknown commitlog version 7ERROR [main] 2020-05-07 01:13:26,798 CassandraDaemon.java:749 - Exception encountered during startupjava.lang.IllegalStateException: Unknown commitlog version 7 at org.apache.cassandra.db.commitlog.CommitLogDescriptor.getMessagingVersion(CommitLogDescriptor.java:227) ~[main/:na] at org.apache.cassandra.db.commitlog.CommitLogReader.shouldSkipSegmentId(CommitLogReader.java:276) ~[main/:na] at org.apache.cassandra.db.commitlog.CommitLogReader.readCommitLogSegment(CommitLogReader.java:220) ~[main/:na] at org.apache.cassandra.db.commitlog.CommitLogReader.readAllFiles(CommitLogReader.java:132) ~[main/:na] at org.apache.cassandra.db.commitlog.CommitLogReplayer.replayFiles(CommitLogReplayer.java:137) ~[main/:na] at org.apache.cassandra.db.commitlog.CommitLog.recoverFiles(CommitLog.java:167) ~[main/:na] at org.apache.cassandra.db.commitlog.CommitLog.recoverSegmentsOnDisk(CommitLog.java:148) ~[main/:na] at org.apache.cassandra.service.CassandraDaemon.setup(CassandraDaemon.java:324) [main/:na] at org.apache.cassandra.service.CassandraDaemon.activate(CassandraDaemon.java:620) [main/:na] at org.apache.cassandra.service.CassandraDaemon.main(CassandraDaemon.java:732) [main/:na]

We manually deleted all new version commit logs and it worked. However, it's not obvious whether manually deleting these commit logs are safe. 

• And when we ran the "ALTER ... DROP COMPACT STORAGE" command on old version node as the error message suggested, non-existent columns appeared with null value.

original table:

 key                    | C0                                                                     | C1                                                                     | C2                                                                     | C3                                                                     | C4
------------------------+------------------------------------------------------------------------+------------------------------------------------------------------------+------------------------------------------------------------------------+------------------------------------------------------------------------+------------------------------------------------------------------------
 0x36384b30374f33314c30 | 0x3ab9b990a5dbbe3dbf0354d35bc32122ab09f963a9a29b04434fac1ab16af8992ed0 | 0xdd9f9d6c207f3720d01cc2f9ab4e83a5c59226c779de8cf79cfe8f1a6c5a092fbee4 | 0x9116a0bb80af7f73d803ceeb78f8aada9555e78936b26678202f6fb1f4993df92089 | 0xbde6fe3df138b0428bcec8ad4ab0dcf256b5bd57c2b6ac42523913522d78f710ee98 | 0x63f1cac68e33a7375a8370fbf300dcc70f76052c3ce84588f3b7d69ef6bc470db986
 0x30374b37384e364c3531 | 0xce7990de95e1516101cbbd6ca3bdc2819e799c8f9b1bfd1b08aa1d1edf09dd409b7d | 0xc99b2076286ee4d4be742508653ed1178fb04192ae192d31745235e57dead6bf7f45 | 0xb492df82f1f2055af30694f135d3c99b0eac4e8d7d4d8e8b2d8ce49a9a3e50e3c63c | 0xc42bcb9b1a215a8d9629887bee918437fd580f0d15c48e1402fe11f6caab069e95aa | 0x329f193b16024ea72ace70571848e56b36496a05896454d13e1696c5c21053b5bcbb

After dropping compact storage:

 key                    | column1 | C0                                                                     | C1                                                                     | C2                                                                     | C3                                                                     | C4                                                                     | value
------------------------+---------+------------------------------------------------------------------------+------------------------------------------------------------------------+------------------------------------------------------------------------+------------------------------------------------------------------------+------------------------------------------------------------------------+-------
 0x36384b30374f33314c30 |    null | 0x3ab9b990a5dbbe3dbf0354d35bc32122ab09f963a9a29b04434fac1ab16af8992ed0 | 0xdd9f9d6c207f3720d01cc2f9ab4e83a5c59226c779de8cf79cfe8f1a6c5a092fbee4 | 0x9116a0bb80af7f73d803ceeb78f8aada9555e78936b26678202f6fb1f4993df92089 | 0xbde6fe3df138b0428bcec8ad4ab0dcf256b5bd57c2b6ac42523913522d78f710ee98 | 0x63f1cac68e33a7375a8370fbf300dcc70f76052c3ce84588f3b7d69ef6bc470db986 |  null
 0x30374b37384e364c3531 |    null | 0xce7990de95e1516101cbbd6ca3bdc2819e799c8f9b1bfd1b08aa1d1edf09dd409b7d | 0xc99b2076286ee4d4be742508653ed1178fb04192ae192d31745235e57dead6bf7f45 | 0xb492df82f1f2055af30694f135d3c99b0eac4e8d7d4d8e8b2d8ce49a9a3e50e3c63c | 0xc42bcb9b1a215a8d9629887bee918437fd580f0d15c48e1402fe11f6caab069e95aa | 0x329f193b16024ea72ace70571848e56b36496a05896454d13e1696c5c21053b5bcbb |  null

•  When we ran stress tool to read, IOExceptions were thrown, while it worked well on old version nodes
• The error message is shown below:

java.io.IOException: Operation x0 on key(s) [4d32304c33314e374c30]: Data returned was not validatedjava.io.IOException: Operation x0 on key(s) [4d32304c33314e374c30]: Data returned was not validated
 at org.apache.cassandra.stress.Operation.error(Operation.java:127) at org.apache.cassandra.stress.Operation.timeWithRetry(Operation.java:105) at org.apache.cassandra.stress.operations.predefined.CqlOperation.run(CqlOperation.java:91) at org.apache.cassandra.stress.operations.predefined.CqlOperation.run(CqlOperation.java:99) at org.apache.cassandra.stress.operations.predefined.CqlOperation.run(CqlOperation.java:248) at org.apache.cassandra.stress.StressAction$Consumer.run(StressAction.java:466)java.io.IOException: Operation x0 on key(s) [4d304f334b3432353231]: Data returned was not validated
 at org.apache.cassandra.stress.Operation.error(Operation.java:127) at org.apache.cassandra.stress.Operation.timeWithRetry(Operation.java:105) at org.apache.cassandra.stress.operations.predefined.CqlOperation.run(CqlOperation.java:91) at org.apache.cassandra.stress.operations.predefined.CqlOperation.run(CqlOperation.java:99) at org.apache.cassandra.stress.operations.predefined.CqlOperation.run(CqlOperation.java:248) at org.apache.cassandra.stress.StressAction$Consumer.run(StressAction.java:466)

 

 

Issue Links

• Is contained by: CASSANDRA-15811 Improve DROP COMPACT STORAGE