Video Screencast Help

SRL log volumes in a VVR configuration.

Created: 03 Feb 2014 • Updated: 04 Feb 2014 | 6 comments
This issue has been solved. See solution.


We are using VVR technology for replication from our PROD enviornment ( 2 node solaris ldom cluster) to our DR enviornment (1 node solaris ldom cluster) with Geo-Cluster configured. 

My query is regarding the SRL log volume configuration.  What is the best practice for configuring the SRL log volume for a particular RVG.  Should be it be mirrored, concat or stripe volume, so that it does not affect the application performance.  We have 2 san luns of same size mapped to the servers for the SRL logs, the luns have redundancy (raid 6) at the SAN level.  Also does it matter if we have the DCM logs on the same lun as the SRL log volume?

Thanks in Advance.

Operating Systems:

Comments 6 CommentsJump to latest comment

mikebounds's picture

SRL volume needs to be at least as performant as the data volumes and should be redundant.  In early versions of VVR you needed to stripe at Volume Manager level to get decent performance, but if you have the same raid 6 performance on all your LUNS (for data and SRL), then striping in VM is probably not nessary, but better to do stripe than concat if you have more than one LUN for SRL.

The offical recommendation is to put DCMs on data LUNs, but as with arrays, individual LUNs don't relate to individual disk spindles, it doesn't really make much difference if you put DCMs on SRL LUNs


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

Gaurav Sangamnerkar's picture


Yes, SRL volume needs to have optimum write performance & equal set of mirroring for redundancy. The same is explained in documentation as well

From a VxVM perspective, the SRL is just another volume. Because all writes are written to the SRL first, it is important for the SRL to have optimum write performance. This means all performance techniques used to increase write performance of a volume apply to the SRL. For most implementations, the SRL is striped across multiple drives for write performance, and mirrored to an equal set of drives for protection.

If there is already a RAID 6 implemented at Storage level, I would recommend to go for mirroring.

I would agree that it shouldn't really matter to have DCM luns on same lun as SRL.


PS: If you are happy with the answer provided, please mark the post as solution. You can do so by clicking link "Mark as Solution" below the answer provided.

unix_san's picture


Can you also tell me how exactly is the data written in the SRL log volume in case of a mirror.  Does vxvm write 2 copies of data on to the SRL volume in case of a mirror to get the write acknowledgement.  Also doesn't this have an overhead on the overall write performance.

mikebounds's picture

The SRL is a volume and so each write is sent to SRL volume and data volume.  As with data volumes, if the SRL is mirrored using Veritas Volume Manager then the writes will be written in parallel to each mirror.

Mirroring has very little impact on peformance as the writes are sent in parallel.

Impact of on peformance in using VVR depends on whether you are in async or sync mode:

In async mode, the write is acknowledged when it has been written to the SRL (doesn't have to be written to data volumes for acknowledgement ), so if your SRL is as performant as your data volumes, then there should be little impact on performance (there is about 5% - 10% additional writes for SRL headers), except if your bandwidth is not sufficent so that VVR gets a long way behind as then VVR will need to read from SRL to catch-up, so reading and writing to the SRL volumes can impact performance. If VVR is a little behind (less than 100MB) then catch-up is done from memory so performance of SRL is not effected.

In Sync mode, the write is acknowledged when it has been written to SRL and there has been a network acknowledgement from the secondary that it has received the write, so here the latency is effected.


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