Video Screencast Help
Symantec to Separate Into Two Focused, Industry-Leading Technology Companies. Learn more.

Storage foundation for Windows bandwidth issue..

Created: 23 May 2013 | 5 comments

Hi,

We  have the following license:-

1.VRTS STORAGE FOUNDATION 6.0 WIN FOR OS TIER STANDARD EDITION PER SERVER STD LIC EXPRESS BAND S

2.VRTS STORAGE FOUNDATION OPTION VOLUME REPLICATOR 6.0 WIN FOR OS TIER STANDARD EDITION PER SERVER STD LIC EXPRESS BAND S

We the following infra:-

1. OS - Windows 2008 R2 Ent.

2. Application - Oracle database 11.2g.(Without RAC)

3. Hardware - RAM - 24GB; HDD - C:, D:, E:.

4. HDD - RAID configured in local system.

5. Oracle application is installed in C Driave and Database is in D Drive.

6. Bandwidth - 24mbps/18-20ms (Shared)

I am new in storage foundation.

Please, help me with my query.

1. Will the above license is enough to replicate the drives.

2. Oracle is in D drive will the software able to replicate the drive.

3. Is the shared bandwidth is enough to replicate in real time.

4. Is there any thing I need to check before installations.

5. Any por and corns ...(if any).

Please, provide the doc links for the same.

Thanks.

 

Operating Systems:

Comments 5 CommentsJump to latest comment

mikebounds's picture

See answers:

 

1. Will the above license is enough to replicate the drives.

Yes

2. Oracle is in D drive will the software able to replicate the drive.

Only if D: is a volume in a SFW cluter diskgroup.  If D: is a partition on a basic disk, it cannot be replicated

3. Is the shared bandwidth is enough to replicate in real time.
This depends on how much your database writes - Use Vradvisor to find out

4. Is there any thing I need to check before installations.

Check you have allocated space for the SRL (storage replicator log) and DCM (Data Change Maps which are tiny so you only need a few MB for these)

5. Any por and corns ...(if any).

If D: is on a disk on the same controller as the boot disk, then you need to use "vxclus UseSystemBus ON” so you can create a cluster diskgroup

I would recommend putting temporary DB spaces on a separate volume (only seperate volume - no need for seperate LUN) which you exclude from replication - this will reduce bandwidth requirements and time to catch-up if SRL fills

 

Mike

UK Symantec Consultant in VCS, GCO, SF, VVR, VxAT on Solaris, AIX, HP-ux, Linux & Windows

If this post has answered your question then please click on "Mark as solution" link below

Wally_Heim's picture

I don't think this Mrinal is using a cluster.  There is nothing in the original post that would indicate a cluster is being used.  If this is not a cluster then the diskgroup needs to be a Secondary diskgroup and not a Clustered diskgroup.

 

Also the vxclus command mentioned by Mike would not be needed if this is not a cluster.

 

My main concern here is the location of the D:\ drive letter.  Is it on a disk of its own or is it just a partition on the same disk as the OS C:\ drive.  VVR will not able to replicate it if it is a partition on the same drive that the OS C:\ drive is on. 

In addition to this you will also need additional diskspace to use for the replicator log.  Most configurations use a replicator log that is about 20-33% the size of the volume being replicated.  But as Mike mentioned the vradvisor utility will help you size the replicator log per your specific requirements.

 

For the bandwidth, you will need to determine how much bandwidth is actually available.  "24mbps shared" typically means that actual bandwidth available for replication is much less than 24mbps.  Have your network people tell you the average and peak usage of the link.  This should give you an idea of what is available for replication to use without impacting normal traffic on the link.  Then as long as Oracle updates the disks at less than this speed, then the bandwidth will be fine.  If Oracle is making changes faster than the available bandwidth then replication will fall behind.

You can use VVR compression to maybe increase the amount of data that VVR can replicate beyond the avaialble bandwidth.  This feature compresses the VVR data packets before sending them on the network.  Some data does not compress very well so VVR will send either the compressed or uncompressed packet depending on which is smaller for the data being sent.  I would recommend using using the lastest CP for 6.0 if you are going to use VVR compression.

Thank you,

Wally

mrinal.sarkar6200's picture

Hi,

The D drive is a RAID Partition and separated from C Drive.

The database size is 16-17 GB and the normal bandwidth is 12-13 MBpas.

Its a Bank's database and the pick hours are :-

a. 4PM to 8 PM

b. 10AM to 1PM.

Please, let me know any thing else information required before installation.

ORACLE is not in cluster or RAC.

Thanks.

 

Certified Symantec Professional.

Work mostly with Netbackup, Backup Exec, Storage Foundation, DLO and System Recovery

Mrinal Sarkar

Howrah, India

________________________

mikebounds's picture

If D drive is a partition on a basic disk you will need to upgrade this to a SFW dynamic diskgroup and create an SRL volume, ideally on a different disk (must be in the same diskgroup).  You then create an RDS which will create DCMs for you.  I would initiate the synchronisation using autosync which should take a couple of hours to sync 16GB at about 20mbps

If it is a Bank, then I assume they want synchrounus replication, rather than async - is this correct?

How far apart are the 2 sites?

Is the normal bandwidth 12-13 megabits per second or MegaBytes per second (which is 96 - 104 megabits per second) and what is the peak bandwidth?

I would also ask the DBAs if they know the temp DB  usage - often they don't - the last customer I asked said temp DB was less than 1% the size of database so was insignificant, but it turned out that 80% of writes went to temp, so excluding temp from replication made a huge difference.  If can get an outage then it should be fairly simple to re-create temp DB spaces on separate volume which is what I would recommend.

Mike

UK Symantec Consultant in VCS, GCO, SF, VVR, VxAT on Solaris, AIX, HP-ux, Linux & Windows

If this post has answered your question then please click on "Mark as solution" link below

mrinal.sarkar6200's picture

Hi,

Sorry for my delay response...

The D drive is a volume in the same RAID partition where the C drive reside...

If it is a Bank, then I assume they want synchrounus replication, rather than async - is this correct?

Yes , they want synchrounus replication.

How far apart are the 2 sites?

Its about 2000 to 2500 KM apart between two site.

No, I have asked the same to the DBA, but he doesn;t know the DB temp file size.

Qery :-

1. Is it possible to do a synchrounus replicationwhen it gets a bandwidth and it wil autometically convert to asynch mode whenthe bandwidth is low.

2. I have provided the below ports.

1710, 8999, 8989, 14144, 49152-65535, 8989, 8199, 2148, 3207, 14150, 5634 RPC port and Anonymous ports

Is these ports are OK or I have give more then required...

We have got these ports form the below symantec links.

http://www.symantec.com/business/support/index?page=content&id=TECH83441

http://www.symantec.com/business/support/index?page=content&id=TECH58431

http://www.symantec.com/business/support/index?page=content&id=TECH17221

 

Thanks

Certified Symantec Professional.

Work mostly with Netbackup, Backup Exec, Storage Foundation, DLO and System Recovery

Mrinal Sarkar

Howrah, India

________________________