An array of errors and the fix...

I am posting this in the hope that Google will spider the post and others
may not have to walk the same path that I have over the last week.
I have a new server with SATA HighPoint RAID 5, 2GB of RAM, 3.2GHz Intel P4,
915GAG motherboard. After a successful installation of Win Server 2003 and
SQL Server 2000 (and service packs) I started to bring over the SQL Server
databases off my old server. I was running some queries to check out the
speed and strange things started to happen. Query connections were being
dropped, tables were becoming corrupted and DBCC was causing SQL Server to
GPF. Some of the errors are seen below. I studied each error as it turned up
but most of the time the problem was left unresolved. I thought it was the
RAID array. I removed it but it still happened. I installed WinXP but it
still happened. I thought it was the network cables but it still happened.
This excellent article
http://support.microsoft.com/default.aspx?scid=kb;en-us;826433 pointed me to
SQLIOStress (this is a top piece of software - I did not know it even
existed - check it out) Sure enough it indicated that there was a problem
with stale reads and dirty writes; something to do with caching. My hard
drives do have 8 meg of cache. However this was ultimately a red herring.
Eventually I discovered using the help of MemTest86+
(http://www.memtest.org) that I had about 12 bits of corrupted RAM towards
the top of the 2GB memory space. I ripped out the dodgy memory and the
problem disappeared. So if you see any of the following error messages try
doing a Memtest first and save yourself a weeks worth of mucking around.
As an aside, I have now Memtested my other computers to find that one of
them has 2 bits of dodgy RAM. If this posting ever helps you then be sure
to drop me an email.
Regards
Dave A (dave @. sigmasolutions.com.au)
These are the errors that were appearing in the event log:
17310 :SqlDumpExceptionHandler: Process 3796 generated fatal exception
c0000005 EXCEPTION_ACCESS_VIOLATION. SQL Server is terminating this process.
Error: 947, Severity: 16, State: 1 Error while closing database 'myDB'
cleanly
Error: 3414, Severity: 21, State: 1 Database 'myDB' (database ID 7) could
not recover. Contact Technical Support.
Error: 3314, Severity: 21, State: 3 Error while undoing logged operation in
database 'myDB'. Error at log record ID (29979:11238:226).
Error: 644, Severity: 21, State: 1 Could not find the index entry for RID
'36ad1720c37a7a19b9500300101802000' in index page (1:168081), index ID 0,
database 'myDB'.
18052 : Error: 3624, Severity: 20, State: 1.
17066 : SQL Server Assertion: File: <recbase.cpp>, line=1378 Failed
Assertion = 'm_offBeginVar < m_SizeRec'.
17066 : SQL Server Assertion: File:
<p:\sql\ntdbms\storeng\drs\include\record.inl>, line=1447 Failed Assertion
= 'm_SizeRec > 0 && m_SizeRec <= MAXDATAROW'.
17066 : SQL Server Assertion: File: <lckmgr.cpp>, line=5201 Failed Assertion
= 'NULL == m_lockList.Head ()'.Thanks for the info Dave. Can you tell us the brands of the server and the
memory as well?
--
Jacco Schalkwijk
SQL Server MVP
"Dave A" <dave@.sigmasolutionsdonotspamme.com.au> wrote in message
news:uT5O2YrCFHA.3376@.TK2MSFTNGP12.phx.gbl...
>I am posting this in the hope that Google will spider the post and others
> may not have to walk the same path that I have over the last week.
> I have a new server with SATA HighPoint RAID 5, 2GB of RAM, 3.2GHz Intel
> P4,
> 915GAG motherboard. After a successful installation of Win Server 2003 and
> SQL Server 2000 (and service packs) I started to bring over the SQL Server
> databases off my old server. I was running some queries to check out the
> speed and strange things started to happen. Query connections were being
> dropped, tables were becoming corrupted and DBCC was causing SQL Server to
> GPF. Some of the errors are seen below. I studied each error as it turned
> up
> but most of the time the problem was left unresolved. I thought it was
> the
> RAID array. I removed it but it still happened. I installed WinXP but it
> still happened. I thought it was the network cables but it still happened.
> This excellent article
> http://support.microsoft.com/default.aspx?scid=kb;en-us;826433 pointed me
> to
> SQLIOStress (this is a top piece of software - I did not know it even
> existed - check it out) Sure enough it indicated that there was a problem
> with stale reads and dirty writes; something to do with caching. My hard
> drives do have 8 meg of cache. However this was ultimately a red herring.
> Eventually I discovered using the help of MemTest86+
> (http://www.memtest.org) that I had about 12 bits of corrupted RAM towards
> the top of the 2GB memory space. I ripped out the dodgy memory and the
> problem disappeared. So if you see any of the following error messages try
> doing a Memtest first and save yourself a weeks worth of mucking around.
> As an aside, I have now Memtested my other computers to find that one of
> them has 2 bits of dodgy RAM. If this posting ever helps you then be
> sure
> to drop me an email.
> Regards
> Dave A (dave @. sigmasolutions.com.au)
> These are the errors that were appearing in the event log:
> 17310 :SqlDumpExceptionHandler: Process 3796 generated fatal exception
> c0000005 EXCEPTION_ACCESS_VIOLATION. SQL Server is terminating this
> process.
> Error: 947, Severity: 16, State: 1 Error while closing database 'myDB'
> cleanly
> Error: 3414, Severity: 21, State: 1 Database 'myDB' (database ID 7) could
> not recover. Contact Technical Support.
> Error: 3314, Severity: 21, State: 3 Error while undoing logged operation
> in
> database 'myDB'. Error at log record ID (29979:11238:226).
> Error: 644, Severity: 21, State: 1 Could not find the index entry for RID
> '36ad1720c37a7a19b9500300101802000' in index page (1:168081), index ID 0,
> database 'myDB'.
> 18052 : Error: 3624, Severity: 20, State: 1.
> 17066 : SQL Server Assertion: File: <recbase.cpp>, line=1378 Failed
> Assertion = 'm_offBeginVar < m_SizeRec'.
> 17066 : SQL Server Assertion: File:
> <p:\sql\ntdbms\storeng\drs\include\record.inl>, line=1447 Failed
> Assertion
> = 'm_SizeRec > 0 && m_SizeRec <= MAXDATAROW'.
> 17066 : SQL Server Assertion: File: <lckmgr.cpp>, line=5201 Failed
> Assertion
> = 'NULL == m_lockList.Head ()'.
>
>|||It was VData RAM but the rest was just bought in a shopping list format from
the local computer shop. I have now replaced the RAM and it is working fine.
Just a bad batch.
"Jacco Schalkwijk" <jacco.please.reply@.to.newsgroups.mvps.org.invalid> wrote
in message news:u7wdU9rCFHA.3100@.TK2MSFTNGP09.phx.gbl...
> Thanks for the info Dave. Can you tell us the brands of the server and the
> memory as well?
> --
> Jacco Schalkwijk
> SQL Server MVP
>
> "Dave A" <dave@.sigmasolutionsdonotspamme.com.au> wrote in message
> news:uT5O2YrCFHA.3376@.TK2MSFTNGP12.phx.gbl...
> >I am posting this in the hope that Google will spider the post and others
> > may not have to walk the same path that I have over the last week.
> >
> > I have a new server with SATA HighPoint RAID 5, 2GB of RAM, 3.2GHz Intel
> > P4,
> > 915GAG motherboard. After a successful installation of Win Server 2003
and
> > SQL Server 2000 (and service packs) I started to bring over the SQL
Server
> > databases off my old server. I was running some queries to check out the
> > speed and strange things started to happen. Query connections were being
> > dropped, tables were becoming corrupted and DBCC was causing SQL Server
to
> > GPF. Some of the errors are seen below. I studied each error as it
turned
> > up
> > but most of the time the problem was left unresolved. I thought it was
> > the
> > RAID array. I removed it but it still happened. I installed WinXP but it
> > still happened. I thought it was the network cables but it still
happened.
> > This excellent article
> > http://support.microsoft.com/default.aspx?scid=kb;en-us;826433 pointed
me
> > to
> > SQLIOStress (this is a top piece of software - I did not know it even
> > existed - check it out) Sure enough it indicated that there was a
problem
> > with stale reads and dirty writes; something to do with caching. My hard
> > drives do have 8 meg of cache. However this was ultimately a red
herring.
> >
> > Eventually I discovered using the help of MemTest86+
> > (http://www.memtest.org) that I had about 12 bits of corrupted RAM
towards
> > the top of the 2GB memory space. I ripped out the dodgy memory and the
> > problem disappeared. So if you see any of the following error messages
try
> > doing a Memtest first and save yourself a weeks worth of mucking around.
> >
> > As an aside, I have now Memtested my other computers to find that one of
> > them has 2 bits of dodgy RAM. If this posting ever helps you then be
> > sure
> > to drop me an email.
> >
> > Regards
> > Dave A (dave @. sigmasolutions.com.au)
> >
> > These are the errors that were appearing in the event log:
> >
> > 17310 :SqlDumpExceptionHandler: Process 3796 generated fatal exception
> > c0000005 EXCEPTION_ACCESS_VIOLATION. SQL Server is terminating this
> > process.
> >
> > Error: 947, Severity: 16, State: 1 Error while closing database 'myDB'
> > cleanly
> >
> > Error: 3414, Severity: 21, State: 1 Database 'myDB' (database ID 7)
could
> > not recover. Contact Technical Support.
> >
> > Error: 3314, Severity: 21, State: 3 Error while undoing logged operation
> > in
> > database 'myDB'. Error at log record ID (29979:11238:226).
> >
> > Error: 644, Severity: 21, State: 1 Could not find the index entry for
RID
> > '36ad1720c37a7a19b9500300101802000' in index page (1:168081), index ID
0,
> > database 'myDB'.
> >
> > 18052 : Error: 3624, Severity: 20, State: 1.
> >
> > 17066 : SQL Server Assertion: File: <recbase.cpp>, line=1378 Failed
> > Assertion = 'm_offBeginVar < m_SizeRec'.
> >
> > 17066 : SQL Server Assertion: File:
> > <p:\sql\ntdbms\storeng\drs\include\record.inl>, line=1447 Failed
> > Assertion
> > = 'm_SizeRec > 0 && m_SizeRec <= MAXDATAROW'.
> >
> > 17066 : SQL Server Assertion: File: <lckmgr.cpp>, line=5201 Failed
> > Assertion
> > = 'NULL == m_lockList.Head ()'.
> >
> >
> >
> >
>