I have a active/passive clustered SQL Server 2005 Enterprise. If the
data file becomes corrupt, I've lost my high availability and because
the database is over 200GB, a restore operation will take hours. Does
anyone have any suggestions for creating a warm stand-by in a
clustered environment? Is this even possible, since the stand-by would
need to become the virutal server resource owner?"Curtis" <curtmorrison@.hotmail.com> wrote in message
news:9d55337f-76a2-49f3-b1c8-18b17d09ab16@.b64g2000hsa.googlegroups.com...
>I have a active/passive clustered SQL Server 2005 Enterprise. If the
> data file becomes corrupt, I've lost my high availability and because
> the database is over 200GB, a restore operation will take hours. Does
> anyone have any suggestions for creating a warm stand-by in a
> clustered environment? Is this even possible, since the stand-by would
> need to become the virutal server resource owner?
Hi
Database corruption will be an issue that is not covered by hardware
clustering, and may require more time for recovery if that does occur. You
can look at log shipping or database mirroring, but for this to take over
you would need to change the client and possibly an active/active cluster.
You can also look at online restores to possibly make the system available
quicker in the event of having to restore. See
http://www.microsoft.com/sql/technologies/highavailability/default.mspx
John
Showing posts with label passive. Show all posts
Showing posts with label passive. Show all posts
Monday, March 12, 2012
Friday, March 9, 2012
High Availability
Hi all,
Our production DB is around 4GB and peaks at around 8000
orders a day. We currently have a active/passive cluster
failover solution. During the holiday season, the quorum
disk for the cluster failed. It unfortunately took a long
time (7-10 hours) to recover. Now, someone is suggesting
that we set up an additional and separate active/passive
cluster, and set up log shipping between the existing
cluster (primary) and the new cluster (secondary).
My concern is that it may take too long to make the
secondary server ready for production. We have
replication running on the server, it could take time to
reconfigure the articles.
What are your concerns over this solution? How did you
set up your server?
http://www.microsoft.com/technet/treeview/default.asp?
url=/TechNet/prodtechnol/sql/reskit/sql2000/part4/c1561.asp
Thanks in advance,
AnnaHi, if you want a high availability solution, you need
spare disk in your storages, even with spare, your Quorum
disk must be in a RAID-1 array.
In your scenario, if your primary server failed, you just
have to change the IP of you secondary server to IP of
your primary server.
And you have to be a nice backup schema.
Cesar Kubo
>--Original Message--
>Hi all,
>Our production DB is around 4GB and peaks at around 8000
>orders a day. We currently have a active/passive cluster
>failover solution. During the holiday season, the quorum
>disk for the cluster failed. It unfortunately took a long
>time (7-10 hours) to recover. Now, someone is suggesting
>that we set up an additional and separate active/passive
>cluster, and set up log shipping between the existing
>cluster (primary) and the new cluster (secondary).
>My concern is that it may take too long to make the
>secondary server ready for production. We have
>replication running on the server, it could take time to
>reconfigure the articles.
>What are your concerns over this solution? How did you
>set up your server?
>
>http://www.microsoft.com/technet/treeview/default.asp?
>url=/TechNet/prodtechnol/sql/reskit/sql2000/part4/c1561.as
p
>Thanks in advance,
>Anna
>.
>
Our production DB is around 4GB and peaks at around 8000
orders a day. We currently have a active/passive cluster
failover solution. During the holiday season, the quorum
disk for the cluster failed. It unfortunately took a long
time (7-10 hours) to recover. Now, someone is suggesting
that we set up an additional and separate active/passive
cluster, and set up log shipping between the existing
cluster (primary) and the new cluster (secondary).
My concern is that it may take too long to make the
secondary server ready for production. We have
replication running on the server, it could take time to
reconfigure the articles.
What are your concerns over this solution? How did you
set up your server?
http://www.microsoft.com/technet/treeview/default.asp?
url=/TechNet/prodtechnol/sql/reskit/sql2000/part4/c1561.asp
Thanks in advance,
AnnaHi, if you want a high availability solution, you need
spare disk in your storages, even with spare, your Quorum
disk must be in a RAID-1 array.
In your scenario, if your primary server failed, you just
have to change the IP of you secondary server to IP of
your primary server.
And you have to be a nice backup schema.
Cesar Kubo
>--Original Message--
>Hi all,
>Our production DB is around 4GB and peaks at around 8000
>orders a day. We currently have a active/passive cluster
>failover solution. During the holiday season, the quorum
>disk for the cluster failed. It unfortunately took a long
>time (7-10 hours) to recover. Now, someone is suggesting
>that we set up an additional and separate active/passive
>cluster, and set up log shipping between the existing
>cluster (primary) and the new cluster (secondary).
>My concern is that it may take too long to make the
>secondary server ready for production. We have
>replication running on the server, it could take time to
>reconfigure the articles.
>What are your concerns over this solution? How did you
>set up your server?
>
>http://www.microsoft.com/technet/treeview/default.asp?
>url=/TechNet/prodtechnol/sql/reskit/sql2000/part4/c1561.as
p
>Thanks in advance,
>Anna
>.
>
High ASYNC_NETWORK_IO value
Hello!
I have been analyzing wait stats using get_waitstats_2005 on ou production S
QLserver (two node Active/Passive SQL Server 2005 64-bit cluster) and notice
d high values of
ASYNC_NETWORK_IO wait type (around 50% of total resource type). From what I
can see, our 1GB Network cards are all right. Has anyone had experience trou
bleshooting this wait type? What could be causing high percentage of this wa
it type?
Thanks,
IgorHi
"imarchenko" wrote:
> Hello!
> I have been analyzing wait stats using get_waitstats_2005 on ou production
SQLserver (two node Active/Passive SQL Server 2005 64-bit cluster) and noti
ced high values of
> ASYNC_NETWORK_IO wait type (around 50% of total resource type). From what
I can see, our 1GB Network cards are all right. Has anyone had experience tr
oubleshooting this wait type? What could be causing high percentage of this
wait type?
> Thanks,
> Igor
50% of not very much is probably not something to worry about! If perfmon
stats such as output queue length show no problems then you are probably ok.
I assume that your 1GB network cards are on a 1GB network that does not have
any bottlenecks? If you can ad an extra card I don't think it would not do
any harm doing that!
John|||Also..
Have you checked for queries that return a large result set? You may want to
use SQL profiler to identify them. If you return an excessive result set the
client may not be able to process this fast enough!
John|||John,
This could well be the case. We have old legacy report application
generating thousands of queries per report.
Thanks,
Igor
"John Bell" <jbellnewsposts@.hotmail.com> wrote in message
news:%23z%23wpltUHHA.192@.TK2MSFTNGP04.phx.gbl...
> Also..
> Have you checked for queries that return a large result set? You may want
> to use SQL profiler to identify them. If you return an excessive result
> set the client may not be able to process this fast enough!
> John
>|||Hi Igor
"imarchenko" wrote:
> John,
> This could well be the case. We have old legacy report application
> generating thousands of queries per report.
> Thanks,
> Igor
Profiling the system will certainly show high I/O and duration queries. Look
for instances where the result set is too wide or too many rows are being
returned that are not being used.
If you can't tune or re-write the queries, maybe a different method of
delivery would be more approprate such as a scheduled report or DTS/SSIS
export. Use of Analysis Service may allow you to process your data in a more
piecemeal way?
John|||John,
AS is out of question at the moment, but we are working on improving this
app.
Thanks again,
Igor
"John Bell" <jbellnewsposts@.hotmail.com> wrote in message
news:A4E11028-D4C0-4845-A347-03C5FEB1BE5D@.microsoft.com...
> Hi Igor
> "imarchenko" wrote:
>
> Profiling the system will certainly show high I/O and duration queries.
> Look
> for instances where the result set is too wide or too many rows are being
> returned that are not being used.
> If you can't tune or re-write the queries, maybe a different method of
> delivery would be more approprate such as a scheduled report or DTS/SSIS
> export. Use of Analysis Service may allow you to process your data in a
> more
> piecemeal way?
> John
>
I have been analyzing wait stats using get_waitstats_2005 on ou production S
QLserver (two node Active/Passive SQL Server 2005 64-bit cluster) and notice
d high values of
ASYNC_NETWORK_IO wait type (around 50% of total resource type). From what I
can see, our 1GB Network cards are all right. Has anyone had experience trou
bleshooting this wait type? What could be causing high percentage of this wa
it type?
Thanks,
IgorHi
"imarchenko" wrote:
> Hello!
> I have been analyzing wait stats using get_waitstats_2005 on ou production
SQLserver (two node Active/Passive SQL Server 2005 64-bit cluster) and noti
ced high values of
> ASYNC_NETWORK_IO wait type (around 50% of total resource type). From what
I can see, our 1GB Network cards are all right. Has anyone had experience tr
oubleshooting this wait type? What could be causing high percentage of this
wait type?
> Thanks,
> Igor
50% of not very much is probably not something to worry about! If perfmon
stats such as output queue length show no problems then you are probably ok.
I assume that your 1GB network cards are on a 1GB network that does not have
any bottlenecks? If you can ad an extra card I don't think it would not do
any harm doing that!
John|||Also..
Have you checked for queries that return a large result set? You may want to
use SQL profiler to identify them. If you return an excessive result set the
client may not be able to process this fast enough!
John|||John,
This could well be the case. We have old legacy report application
generating thousands of queries per report.
Thanks,
Igor
"John Bell" <jbellnewsposts@.hotmail.com> wrote in message
news:%23z%23wpltUHHA.192@.TK2MSFTNGP04.phx.gbl...
> Also..
> Have you checked for queries that return a large result set? You may want
> to use SQL profiler to identify them. If you return an excessive result
> set the client may not be able to process this fast enough!
> John
>|||Hi Igor
"imarchenko" wrote:
> John,
> This could well be the case. We have old legacy report application
> generating thousands of queries per report.
> Thanks,
> Igor
Profiling the system will certainly show high I/O and duration queries. Look
for instances where the result set is too wide or too many rows are being
returned that are not being used.
If you can't tune or re-write the queries, maybe a different method of
delivery would be more approprate such as a scheduled report or DTS/SSIS
export. Use of Analysis Service may allow you to process your data in a more
piecemeal way?
John|||John,
AS is out of question at the moment, but we are working on improving this
app.
Thanks again,
Igor
"John Bell" <jbellnewsposts@.hotmail.com> wrote in message
news:A4E11028-D4C0-4845-A347-03C5FEB1BE5D@.microsoft.com...
> Hi Igor
> "imarchenko" wrote:
>
> Profiling the system will certainly show high I/O and duration queries.
> Look
> for instances where the result set is too wide or too many rows are being
> returned that are not being used.
> If you can't tune or re-write the queries, maybe a different method of
> delivery would be more approprate such as a scheduled report or DTS/SSIS
> export. Use of Analysis Service may allow you to process your data in a
> more
> piecemeal way?
> John
>
High ASYNC_NETWORK_IO value
Hello!
I have been analyzing wait stats using get_waitstats_2005 on ou production SQLserver (two node Active/Passive SQL Server 2005 64-bit cluster) and noticed high values of
ASYNC_NETWORK_IO wait type (around 50% of total resource type). From what I can see, our 1GB Network cards are all right. Has anyone had experience troubleshooting this wait type? What could be causing high percentage of this wait type?
Thanks,
Igor
Hi
"imarchenko" wrote:
> Hello!
> I have been analyzing wait stats using get_waitstats_2005 on ou production SQLserver (two node Active/Passive SQL Server 2005 64-bit cluster) and noticed high values of
> ASYNC_NETWORK_IO wait type (around 50% of total resource type). From what I can see, our 1GB Network cards are all right. Has anyone had experience troubleshooting this wait type? What could be causing high percentage of this wait type?
> Thanks,
> Igor
50% of not very much is probably not something to worry about! If perfmon
stats such as output queue length show no problems then you are probably ok.
I assume that your 1GB network cards are on a 1GB network that does not have
any bottlenecks? If you can ad an extra card I don't think it would not do
any harm doing that!
John
|||Also..
Have you checked for queries that return a large result set? You may want to
use SQL profiler to identify them. If you return an excessive result set the
client may not be able to process this fast enough!
John
|||John,
This could well be the case. We have old legacy report application
generating thousands of queries per report.
Thanks,
Igor
"John Bell" <jbellnewsposts@.hotmail.com> wrote in message
news:%23z%23wpltUHHA.192@.TK2MSFTNGP04.phx.gbl...
> Also..
> Have you checked for queries that return a large result set? You may want
> to use SQL profiler to identify them. If you return an excessive result
> set the client may not be able to process this fast enough!
> John
>
|||Hi Igor
"imarchenko" wrote:
> John,
> This could well be the case. We have old legacy report application
> generating thousands of queries per report.
> Thanks,
> Igor
Profiling the system will certainly show high I/O and duration queries. Look
for instances where the result set is too wide or too many rows are being
returned that are not being used.
If you can't tune or re-write the queries, maybe a different method of
delivery would be more approprate such as a scheduled report or DTS/SSIS
export. Use of Analysis Service may allow you to process your data in a more
piecemeal way?
John
|||John,
AS is out of question at the moment, but we are working on improving this
app.
Thanks again,
Igor
"John Bell" <jbellnewsposts@.hotmail.com> wrote in message
news:A4E11028-D4C0-4845-A347-03C5FEB1BE5D@.microsoft.com...
> Hi Igor
> "imarchenko" wrote:
>
> Profiling the system will certainly show high I/O and duration queries.
> Look
> for instances where the result set is too wide or too many rows are being
> returned that are not being used.
> If you can't tune or re-write the queries, maybe a different method of
> delivery would be more approprate such as a scheduled report or DTS/SSIS
> export. Use of Analysis Service may allow you to process your data in a
> more
> piecemeal way?
> John
>
I have been analyzing wait stats using get_waitstats_2005 on ou production SQLserver (two node Active/Passive SQL Server 2005 64-bit cluster) and noticed high values of
ASYNC_NETWORK_IO wait type (around 50% of total resource type). From what I can see, our 1GB Network cards are all right. Has anyone had experience troubleshooting this wait type? What could be causing high percentage of this wait type?
Thanks,
Igor
Hi
"imarchenko" wrote:
> Hello!
> I have been analyzing wait stats using get_waitstats_2005 on ou production SQLserver (two node Active/Passive SQL Server 2005 64-bit cluster) and noticed high values of
> ASYNC_NETWORK_IO wait type (around 50% of total resource type). From what I can see, our 1GB Network cards are all right. Has anyone had experience troubleshooting this wait type? What could be causing high percentage of this wait type?
> Thanks,
> Igor
50% of not very much is probably not something to worry about! If perfmon
stats such as output queue length show no problems then you are probably ok.
I assume that your 1GB network cards are on a 1GB network that does not have
any bottlenecks? If you can ad an extra card I don't think it would not do
any harm doing that!
John
|||Also..
Have you checked for queries that return a large result set? You may want to
use SQL profiler to identify them. If you return an excessive result set the
client may not be able to process this fast enough!
John
|||John,
This could well be the case. We have old legacy report application
generating thousands of queries per report.
Thanks,
Igor
"John Bell" <jbellnewsposts@.hotmail.com> wrote in message
news:%23z%23wpltUHHA.192@.TK2MSFTNGP04.phx.gbl...
> Also..
> Have you checked for queries that return a large result set? You may want
> to use SQL profiler to identify them. If you return an excessive result
> set the client may not be able to process this fast enough!
> John
>
|||Hi Igor
"imarchenko" wrote:
> John,
> This could well be the case. We have old legacy report application
> generating thousands of queries per report.
> Thanks,
> Igor
Profiling the system will certainly show high I/O and duration queries. Look
for instances where the result set is too wide or too many rows are being
returned that are not being used.
If you can't tune or re-write the queries, maybe a different method of
delivery would be more approprate such as a scheduled report or DTS/SSIS
export. Use of Analysis Service may allow you to process your data in a more
piecemeal way?
John
|||John,
AS is out of question at the moment, but we are working on improving this
app.
Thanks again,
Igor
"John Bell" <jbellnewsposts@.hotmail.com> wrote in message
news:A4E11028-D4C0-4845-A347-03C5FEB1BE5D@.microsoft.com...
> Hi Igor
> "imarchenko" wrote:
>
> Profiling the system will certainly show high I/O and duration queries.
> Look
> for instances where the result set is too wide or too many rows are being
> returned that are not being used.
> If you can't tune or re-write the queries, maybe a different method of
> delivery would be more approprate such as a scheduled report or DTS/SSIS
> export. Use of Analysis Service may allow you to process your data in a
> more
> piecemeal way?
> John
>
Subscribe to:
Posts (Atom)