November 15, 2011 at 2:25 am
Hello,
I have set a snapshot replication to run once daily, it gets applied to 2 subscriber DB which is on
S S 2005 & S S 2008.
For 4.4 gigs data, it takes 6 mins for the distribution job to apply a snapshot completely to a 2005 Subscriber DB & takes 23-29 mins to apply the same snapshot to a 2008 subscriber DB.
I am surprise to see such a big difference in the timings.
Can anyone tell me the best way to improve this, also tell me what checks I need to make to find the route cause of this issue?
Thanks
Asir
November 15, 2011 at 7:22 am
Are there any network issues between the pub and the 2008 box? Do you see any network IO waits?
Are the two subscribers having the same HW configuration?
-Roy
November 16, 2011 at 2:37 am
Are there any network issues between the pub and the 2008 box? Do you see any network IO waits?
Are the two subscribers having the same HW configuration?
-Roy
++++++++++++++++++++++++++++++++++++++++++
Hi Roy,
Below are the answers to your questions, please let me know your thoughts.
Hardware configuration --- Hardware configuration is same the only difference is the RAM,OS & System Type.
On 2005 there is only 2 gigs ram & runs on 32 bit version of Windows server 2003.
Whereas 2008 has 4 gigs ram & runs on a 64 bit version of Windows Server 2008 R2.
Network I/O wait --- The I/O wait time was ranging from 450-800 ms/sec while the snapshot was being applied. Most of the time the I/O remained on 600-700 ms/sec.
I monitored today as well & it took 30 mins to apply the snap shot on 2008 DB & 6 min to apply on 2005 DB.
I have no idea what I need to do if the I/O is the culprit, would appreciate if you give me detailed guideliness to troubleshoot the issue.
Thanks Roy.
Appreciate your inputs.
Regards
Asir
November 16, 2011 at 5:52 am
You have a better HW on the 2008 server. So we can forget about that. Couple of things you might want to look at is the Disk configuration.
Are both the subscribers Disk configuration the same? What is the setting for the DB's Autogrowth for both the DB's?
Network IO can be a culprit but I do not see much evidence to why the difference is so high. Both the 2005 and 2008 are set up with the same Network configuration? Do you have network card configured the same way?
-Roy
November 17, 2011 at 2:18 am
You have a better HW on the 2008 server. So we can forget about that. Couple of things you might want to look at is the Disk configuration.
Are both the subscribers Disk configuration the same? What is the setting for the DB's Autogrowth for both the DB's?
Network IO can be a culprit but I do not see much evidence to why the difference is so high. Both the 2005 and 2008 are set up with the same Network configuration? Do you have network card configured the same way?
-Roy
++++++++++++++++++
Both DB have unrestricted growth for MDF & restricted growth to 2 gigs for LDF.
I checked the Network Card for both the server, both are running on "VMware PCI Ethernet Adapter" & have same configuration.
November 18, 2011 at 5:24 am
There are some great queries done by Glenn Berry. Check for his Super DMV queries. It will give you an idea of what are the bottle necks on the SQL 2008 server. That will help analyze the situation.
-Roy
Viewing 6 posts - 1 through 5 (of 5 total)
You must be logged in to reply to this topic. Login to reply