|
|
| Brandon Chin 2005-04-25, 8:25 pm |
| Hi All. We have a rather large database (250 GB) which we need to run DBCC
checks on. We do regularly run checkstorage but as we know that is not
enough. We tried to run checkdb or checktable and checkalloc on the db but
it took too long (around 14days for total duration) for our server.
With this duration we increased the 16K I/O pool to 1GB and its APF to 90
percent and decreased it when it finished. But that still took a long time.
Is there any other way that this process can be sped up?
Any comment is greatly appreciated.
Brandon
| |
| Jeffrey R. Garbus 2005-04-26, 3:25 am |
| Sybase engineering has told me that they think checkstorage is adequate to
meet all needs.
That said, I tend to stagger things. Run a few tables or indexes per evening
/ batch window. Alternatively, run it on your test / backup box so that it
doesn't impact production.
"Brandon Chin" <Brandon. chin@btfinancialgrou
p.com> wrote in message
news:426d9b89@forums
-2-dub...
> Hi All. We have a rather large database (250 GB) which we need to run DBCC
> checks on. We do regularly run checkstorage but as we know that is not
> enough. We tried to run checkdb or checktable and checkalloc on the db but
> it took too long (around 14days for total duration) for our server.
>
> With this duration we increased the 16K I/O pool to 1GB and its APF to 90
> percent and decreased it when it finished. But that still took a long
time.
>
> Is there any other way that this process can be sped up?
> Any comment is greatly appreciated.
>
> Brandon
>
>
| |
| mpeppler@peppler.org 2005-04-26, 7:24 am |
| > Hi All. We have a rather large database (250 GB) which we
> need to run DBCC checks on. We do regularly run
> checkstorage but as we know that is not enough.
I agree with Jeff - in my experience checkstorage is
sufficient, as long as you check the error reports to see if
you have any persistent errors.
Michael
| |
| Jeffrey R. Garbus 2005-04-29, 11:26 am |
| What, check the output of our preventive maintenance jobs to see if they
were successful? What a novel concept! *grin*
Really though, thanks for mentioning that mike. From a practical standpoint,
it's best if you automate the run of the checks as well as the scan of the
reporting output.
Jeff
<mpeppler@peppler.org> wrote in message
news:426e045c.1f96.1681692777@sybase.com...
>
> I agree with Jeff - in my experience checkstorage is
> sufficient, as long as you check the error reports to see if
> you have any persistent errors.
>
> Michael
| |
| Brandon Chin 2005-05-01, 8:27 pm |
| Thanks guys for responding. We are sort of doing that anyway but just
thought that there might be a better way.
cheers
Brandon
"Jeffrey R. Garbus" <jeff@soaringeagle.biz> wrote in message
news:427250a4$1@foru
ms-2-dub...
> What, check the output of our preventive maintenance jobs to see if they
> were successful? What a novel concept! *grin*
>
> Really though, thanks for mentioning that mike. From a practical
> standpoint,
> it's best if you automate the run of the checks as well as the scan of the
> reporting output.
>
> Jeff
> <mpeppler@peppler.org> wrote in message
> news:426e045c.1f96.1681692777@sybase.com...
>
>
| |
| Bret Halford 2005-05-04, 11:24 am |
| Someone in engineering may have expressed that opinion, but I for one
heartily disagree. And it seems clearly obvious - checkstorage does not
check index consistency.
-bret
"Jeffrey R. Garbus" wrote:
[color=darkred]
> Sybase engineering has told me that they think checkstorage is adequate to
> meet all needs.
>
> That said, I tend to stagger things. Run a few tables or indexes per evening
> / batch window. Alternatively, run it on your test / backup box so that it
> doesn't impact production.
>
> "Brandon Chin" <Brandon. chin@btfinancialgrou
p.com> wrote in message
> news:426d9b89@forums
-2-dub...
> time.
|
|
|
|