[
https://issues.apache.org/jira/browse/DERBY-5450?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13123561#comment-13123561
]
Mike Matrigali commented on DERBY-5450:
---------------------------------------
i will take a look on monday if noone else gets to it. If the db is a of a
reasonable size can you also attach it to the issue. if we are lucky the old
transaction logs may help tell what was going on with that page. If not the
next step would be to see if you can reproduce on that same environment with
the following set:
derby.storage.keepTransactionLog=true
This will make the system keep all transaction logs, rather than deleting them.
Thus it gives a complete history to all updates that led to the error. It
will make the db bigger so you will need more disk space for the log directory
of the database. you can estimate how much
you will need by looking at the log directory of the db that had the problem
and see the number and size of the last log file there and
multiply to give approximately what all the log files will take. IT is not
guaranteed as the log files can vary in size but at least a good
minimum.
> in nstest: XSLA6: Cannot recover from database & and 40XT4 errors in nstest
> (sane jars) after ASSERT FAILED
> org.apache.derby.shared.common.sanity.AssertFailure: ASSERT FAILED
> inconsistency in space management during insert
> ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
>
> Key: DERBY-5450
> URL: https://issues.apache.org/jira/browse/DERBY-5450
> Project: Derby
> Issue Type: Bug
> Affects Versions: 10.8.2.2
> Environment: SUSE Linux 10, ibm 1.6 SR9 FP1. 2 CPU machine, but
> hyperthreading on
> Reporter: Myrna van Lunteren
> Attachments: d5450_derbylog.jar
>
>
> With the 10.8 (10.8.2.2) tree sync-ed up to revision: 1179754, I ran NsTest
> Embedded with sane jars, but no derby.properties file.
> I saw the following AssertFailure:
> org.apache.derby.shared.common.sanity.AssertFailure: ASSERT FAILED
> inconsistency in space management during insert: slot = 146
> getSlotOffset(slot) = 3206 dataWritten = 19 freeSpace = -15 firstFreeByte =
> 3149 page = ---------------------------------------------------
> page id: Page(378,Container(0, 1153)) Overflow: false PageVersion: 895
> SlotsInUse: 159 DeletedRowCount: 93 PageStatus: 1 NextId: 317 firstFreeByte:
> 3149 freeSpace: -15 totalSpace: 4028 spareSpace: 0% minimumRecordSize : 1
> PageSize: 4096
> This was followed by:
> ------------ BEGIN SHUTDOWN ERROR STACK -------------
> ERROR XSLA6: Cannot recover the database.
> and somewhat later:
> ERROR 40XT4: An attempt was made to close a transaction that was still
> active. The transaction has been aborted.
> But after a while we get back to the errors always seen with this test (ERROR
> 22003 and XBM06).
> I will attach the derby.log, in case someone likes to look at puzzling
> behavior.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
