SDS Replication and VMware VMFS
Posted in 2016
Topics: High Availability & Replication, Platform-Specific Issues
Hello, Ihac who wants to implement SDS replication using two Vmware Linux virtual machines. Does VMFS support SDS replication? Thanks a lot for your answers. JOHN.
Maybe I was not very clear making the question.What I wanted to ask is the following: When configuring SDS servers as ESXi virtual servers can we use a VMware virtual disk as the shared disk or is the only supported way to pass through a SAN or iSCSI LUN to both SDS nodes?
I think you didn't get an answer because no one knew one. Are the ESXi virtual disks shareable? If yes, then I suppose they will work. Art Art S. Kagel, President and Principal Consultant ASK Database Management www.askdbmgt.com Blog: http://informix-myview.blogspot.com/ Disclaimer: Please keep in mind that my own opinions are my own opinions and do not reflect on the IIUG, nor any other organization with which I am associated either explicitly, implicitly, or by inference. Neither do those opinions reflect those of other individuals affiliated with any entity with which I am affiliated nor those of the entities themselves. On Tue, Mar 15, 2016 at 2:10 PM, JUAN ROCA <jluis.roca@hotmail.com> wrote: > Maybe I was not very clear making the question.What I wanted to ask is the > following: > > When configuring SDS servers as ESXi virtual servers can we use a VMware > virtual disk as the shared disk or is the only supported way to pass > through a > SAN or iSCSI LUN to both SDS nodes? > > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. > > --001a1141f23c6ea83b052e1e8d39
blockquote, div.yahoo_quoted { margin-left: 0 !important; border-left:1px #715FFA solid !important; padding-left:1ex !important; background-color:white !important; } I can't answer regarding the specific shared disk solution. However, I can state what the actual requirement is for any successful implementation of SDS. The fundamental requirement is that when an IO is completed on the primary of the SDS cluster, that the IO be immediately readable on the secondary nodes. There can be no buffering on the primary, nor can there be cash reads on the secondary. The reason for this is that if this requirement is not met, then the secondary will not be able to read the current version of the page, and that will cause data corruption in the buffer pool on the secondary, leading to a failure. That is the reason that cross mounted files will not work, for instance. There is simply too much OS buffering with basic cross mounted files - and that will result in failure. If you can confirm that your configuration will meet this requirement, then it should be ok. If it does not, then it will probably fail. Sent from Yahoo Mail for iPad On Tuesday, March 15, 2016, 3:19 PM, JUAN ROCA <jluis.roca@hotmail.com> wrote: Maybe I was not very clear making the question.What I wanted to ask is the following: When configuring SDS servers as ESXi virtual servers can we use a VMware virtual disk as the shared disk or is the only supported way to pass through a SAN or iSCSI LUN to both SDS nodes? ******************************************************************************* Forum Note: Use "Reply" to post a response in the discussion forum.
A starting point for further considerations, maybe: =20 http://www.dcig.com/2013/03/vmware-vmfs-multi-writer-flag.html They definitely call it a cluster file system .... From: "Madison Pruet" <madison=5Fpruet@yahoo.com> To: ids@iiug.org Date: 16.03.2016 05:51 Subject: Re: SDS Replication and VMware VMFS [36762] Sent by: ids-bounces@iiug.org blockquote, div.yahoo=5Fquoted { margin-left: 0 !important; border-left:1px= =20 #715FFA solid !important; padding-left:1ex !important;=20 background-color:white=20 !important; } I can't answer regarding the specific shared disk solution.=20 However, I can state what the actual requirement is for any successful=20 implementation of SDS. The fundamental requirement is that when an IO is=20 completed on the primary of the SDS cluster, that the IO be immediately=20 readable on the secondary nodes. There can be no buffering on the primary, = nor=20 can there be cash reads on the secondary. The reason for this is that if=20 this=20 requirement is not met, then the secondary will not be able to read the=20 current version of the page, and that will cause data corruption in the=20 buffer=20 pool on the secondary, leading to a failure. That is the reason that cross = mounted files will not work, for instance. There is simply too much OS=20 buffering with basic cross mounted files - and that will result in=20 failure.=20 If you can confirm that your configuration will meet this requirement,=20 then it=20 should be ok. If it does not, then it will probably fail.=20 Sent from Yahoo Mail for iPad=20 On Tuesday, March 15, 2016, 3:19 PM, JUAN ROCA <jluis.roca@hotmail.com>=20 wrote:=20 Maybe I was not very clear making the question.What I wanted to ask is the = following:=20 When configuring SDS servers as ESXi virtual servers can we use a VMware=20 virtual disk as the shared disk or is the only supported way to pass=20 through a=20 SAN or iSCSI LUN to both SDS nodes?=20 ***************************************************************************= ****=20 Forum Note: Use "Reply" to post a response in the discussion forum.=20 ***************************************************************************= ****=20 Forum Note: Use "Reply" to post a response in the discussion forum.=20