I am working with an uncommon situation. I have a distributed application which reads data from flat files on a disk (an old-school "database"). Previously, the "database" would be on N computers as dedicated hard drives. (These are windows computers FYI) We are looking to make these computers diskless, and using iSCSI as the disk technology. I am working on a proof-of-concept using tgt, and seem to be having some limited success. I have mounted the "database" LUN as read-only on one system, and read-write on another system (as a different LUN). I know that this is a non-standard way to use iSCSI, but there are some advantages to this strategy. Only one copy of data would need to be maintained. Since the amount of data being transferred is relatively high, we were concerned with the additional overhead of using CIFS/SAMBA. My success has revolved around setting allow-in-use to "yes" (otherwise only one target will have the LUN tied to the backing store). Now, when I place a file on the database disk using the read/write LUN, it doesn't show up on the read-only LUN until it is unmounted and re-mounted on the client system. I can see a situation where the iSCSI standard might make this to be the correct behavior (since it's read-only, you'll never have the contents change, right?) Is there a way to change this behavior? Any other oddities that I may expect to run into? Thanks -- To unsubscribe from this list: send the line "unsubscribe stgt" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
