Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
Hi, Thanks for your reply Felix. The problem with our application is that it should history data of devices that continiously generate data, So the files that application is using them are open for long time (for example one day) and modify quickly (for example each seconds). Also application, buffers data and flush the buffer periodically into files. So the questions are: 1- How long is resynchronization time, after a node become up after outage or split brain? 2- Does DRBD works fine even when the files are open and modified quickly? 3- In case of using VM, what happens to application buffer befor flushing? Does the VM disk cache also will be replicated? any idea will be appreciated. Regards On Mon, May 2, 2011 at 3:58 PM, Felix Frank <ff at mpexnet.de> wrote: > On 05/02/2011 01:10 PM, Meisam Mohammadkhani wrote: > > Hi All, > > > > I'm new to DRBD. I'm searching around a solution for our enterprise > > application that is responsible to save(and manipulate) historical data > > of industrial devices. Now, we have two stations that works like hot > > redundant of each other. Our challenge is in case of failure. Our > > application is running on two totally independent machines and so each > > one has its own disk (it can be scale into more machines and even in WAN > > connectivity in some projects, but we are following spec that I said > > before for now). Applications are writing into many files continiously. > > We are searching around a solution like a "high available transparent > > file system" that makes the fault transparent to the application, so in > > case of fault, redundant machine still can access the files even the > > master machine is down (replica issue or such a thing). Is there any > > feature of DRBD which can help us in this? Also this is notable that our > > application is written in .net so requires windows platform, so we > > should use virtualization if we wanna use DRBD. > > > Hi, > > this is actually a very common use case for DRBD: > > Have DRBD sync a disk partition between your machines. This partition > contains the filesystem of a virtual machine. When the active host > fails, start the VM on the secondary host and continue working. > > Note that you will suffer performance penalties for a) using > virtualization and b) syncing with DRBD. > If performance is not crucial to your application, there are hardly any > downsides to this setup. > > Regards, > Felix > -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.linbit.com/pipermail/drbd-user/attachments/20110502/9962b6cb/attachment.htm>