db-derby-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Stan Bradbury (JIRA)" <j...@apache.org>
Subject [jira] Commented: (DERBY-2475) invalid checksum
Date Mon, 09 Jul 2007 15:26:04 GMT

    [ https://issues.apache.org/jira/browse/DERBY-2475?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12511156
] 

Stan Bradbury commented on DERBY-2475:
--------------------------------------

These problems could be caused by the environment being used, I/O issues or JVM / memory problems.
 One vexing issue (vexing because it cannot be reproduced reliably and is set by default for
Windows) is the issue write caching being enabled  for disks. I recommend disabling this for
all database disks as it can cause data loss (e.g. corruption) if there is a failure.  For
information on disabling write caching see:   http://support.microsoft.com/kb/259716.
[  NOTE  the warning in RED at the end of this article ]

 I've seen file / store-level corruptions like this caused by environmental factors.  It wouild
be very helpful to get detailed environment information with these reports.  In particular:
   SYSINFO information to identify build information and environment settings
   java -version output to provide JVM specifics (build)
   Description of the processor - SMP?  
   Description of the storage systems - RAM disks?  RAID arrays?
   Local disks (not network mounted disks) being used?

The original report where copying the database to another machine resolves the problem sounds
like the problem was a in-memory corruption (assuming the system was shutdown and the entire
directory tree copied) - possibly a JVM issue?

> invalid checksum
> ----------------
>
>                 Key: DERBY-2475
>                 URL: https://issues.apache.org/jira/browse/DERBY-2475
>             Project: Derby
>          Issue Type: Bug
>          Components: Store
>    Affects Versions: 10.2.2.0
>         Environment: windows xp, notebook, possibly broken hard drive!
>            Reporter: Fabian Merki
>            Priority: Critical
>         Attachments: defekt2.zip
>
>
> one of my customers gets on a notebook the following exception:
> ----------------------------------------------------------------
> 2007-03-17 09:01:53.042 GMT:
>  Booting Derby version The Apache Software Foundation - Apache Derby - 10.2.2.0 - (485682):
instance c013800d-0111-5f21-a736-000002cd2118
> on database directory C:\Programme\merkisoft\kursverwaltung\database\kurs  
> Database Class Loader started - derby.database.classpath=''
> ------------  BEGIN SHUTDOWN ERROR STACK -------------
> ERROR XSDG2: Invalid checksum on Page Page(28,Container(0, 1248)), expected=3'455'715'557,
on-disk version=3'357'396'866, page dump follows: Hex dump
>  
> hex dump => see attachment
>  
> i'm using the embedded version together with hibernate on the notebook.
> when i transfer the db to my pc and execute select * from <all tables> i get no
errors at all.
> we checked the disk about 3 times (with error correction): no result and it still happens.
> when i create a new db on the notebook it happens after a while again.
> the db is about 35 mb. i could put it somewhere for developers (please email me).
>  
> i've seen someone else having a similar problem but i couldn't find a solution yet.
>  
> any help is very welcome!
>  
> ------------------
> hi bryan
> thanks for the quick reply! i tried to answer your questions. i hope it
> helps.
> > Does this happen for this customer over and over?
> > Or did it happen only once?
> yes only that, but it happend 6 times.
> > Is the customer able to provoke the error on demand?
> i tried without success but i keep testing it.
> > Can you relate the exception to any particular action that the customer
> > is taking in your application?
> till now no.
> > If it has happened multiple times, is it always the same page on
> > the same container? Or is the location different each time?
> we installed the app serveral times and i think the following lines are per
> installation (so per installation it happend on the same page):
> org.hibernate.util.JDBCExceptionReporter - Invalid checksum on Page
> Page(28,Container(0, 1248)), expected=3'455'715'557, on-disk
> version=3'357'396'866, page dump follows: Hex dump:
> org.hibernate.util.JDBCExceptionReporter - Invalid checksum on Page
> Page(461,Container(0, 1248)), expected=3'908'279'257, on-disk
> version=137'782'528, page dump follows: Hex dump:
> org.hibernate.util.JDBCExceptionReporter - Invalid checksum on Page
> Page(482,Container(0, 1248)), expected=4'113'528'744, on-disk
> version=44'301'386, page dump follows: Hex dump:
> org.hibernate.util.JDBCExceptionReporter - Invalid checksum on Page
> Page(702,Container(0, 944)), expected=2'598'292'545, on-disk
> version=3'174'535'138, page dump follows: Hex dump:
> org.hibernate.util.JDBCExceptionReporter - Invalid checksum on Page
> Page(2767,Container(0, 944)), expected=2'067'623'629, on-disk
> version=2'180'532'620, page dump follows: Hex dump:
> org.hibernate.util.JDBCExceptionReporter - Invalid checksum on Page
> Page(99,Container(0, 1248)), expected=242'030'294, on-disk
> version=3'389'026'885, page dump follows: Hex dump:
>  > When the exception happens, are there any follow-on symptoms?
> yes, my app crashes ;-((
> i'm not shure how well i reinit the db connection.
> > Or did > the application seem to behave properly after the user shut it
> > down and restarted it? Can the customer later access the data properly
> > on their notebook? Or do you have to move the files to another machine
> > in order to access the data?
> restarts did not help as far as i know. but but when i copied it onto my
> harddrive it was ok.
> > thanks,
> thank you!
> fabian
> thanks
> fabian

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message