Checkpoint - Not running?

Hi folks.
I have a SQL cluster running 2000. My databases are set to simple
recovery. The recovery interval on the server is set to zero.
My logfiles are continually growing on the servers in the cluster -
including the one in tempdb. I do not have long-running transactions,
nor do I have large transactions occurring. The minute I manually do a
checkpoint, the logfiles drop to under 1MB in size. Note that the size
of the logfiles grow to over 70% of the total size allocated in the
file and do cause the logfiles to autogrow.
It was my understanding that a checkpoint will occur if the logfile
size is over 70% of the total allocated size of the file.
Checkpointing is definitely not occurring.
Can anyone offer some suggestions or insight as to why the checkpoint
process seems...Dead?
TIA,
Brian
SP4 Build 8.00.2171 has a fix to this. It sorted the problem for us.
Hot fix/build list
http://www.aspfaq.com/sql2000builds.asp
Paul
<lameduck1000@.hotmail.com> wrote in message
news:1159455084.684715.57620@.m73g2000cwd.googlegro ups.com...
> Hi folks.
> I have a SQL cluster running 2000. My databases are set to simple
> recovery. The recovery interval on the server is set to zero.
> My logfiles are continually growing on the servers in the cluster -
> including the one in tempdb. I do not have long-running transactions,
> nor do I have large transactions occurring. The minute I manually do a
> checkpoint, the logfiles drop to under 1MB in size. Note that the size
> of the logfiles grow to over 70% of the total size allocated in the
> file and do cause the logfiles to autogrow.
> It was my understanding that a checkpoint will occur if the logfile
> size is over 70% of the total allocated size of the file.
> Checkpointing is definitely not occurring.
> Can anyone offer some suggestions or insight as to why the checkpoint
> process seems...Dead?
> TIA,
> Brian
>
|||To verify whether or when a checkpoint is issued, you can turn on trace flag
3502 so that an entry is written to the errorlog when a checkpoint is issued.
Linchi
"lameduck1000@.hotmail.com" wrote:

> Hi folks.
> I have a SQL cluster running 2000. My databases are set to simple
> recovery. The recovery interval on the server is set to zero.
> My logfiles are continually growing on the servers in the cluster -
> including the one in tempdb. I do not have long-running transactions,
> nor do I have large transactions occurring. The minute I manually do a
> checkpoint, the logfiles drop to under 1MB in size. Note that the size
> of the logfiles grow to over 70% of the total size allocated in the
> file and do cause the logfiles to autogrow.
> It was my understanding that a checkpoint will occur if the logfile
> size is over 70% of the total allocated size of the file.
> Checkpointing is definitely not occurring.
> Can anyone offer some suggestions or insight as to why the checkpoint
> process seems...Dead?
> TIA,
> Brian
>