On Jun 18, 2011, at 2:27 PM, Adam Levin wrote: > That's great to know. I've been pretty impressed with what I've heard > about Isilon, and they've been around for a while. > > [...] > Are you doing NDMP backups or any replication?
We're using SyncIQ here, and quite happy with it. Basic setup is dead simple, with various knobs to adjust as needed: - Copy vs. sync - Manual vs. scheduled - Source and destination - Files to match within the source (pattern matching against file metadata) - Snapshot policy for destination - Select cluster nodes to use on source and destination clusters The current version has to crawl the filesystem. Our base runtime (i.e., no changes to replicate) with ~9M files is about 90 minutes. A typical transfer is ~1.5T and takes about two hours start to finish. I'm told the next major release won't need to crawl the filesystem. In a disaster scenario, we just update DNS so that everybody starts mounting the backup cluster. When the production cluster comes back online, we execute a reverse sync, revert the DNS changes, and go back to business as usual. The only gotcha is making sure your production cluster doesn't try to run its scheduled replication before you complete your reverse sync. Ours runs nightly, so it isn't much of a problem. Somebody with a more frequent schedule would have to be a bit more careful though. Regards, Jonathan Jonathan Rozes director, information technology, LAIKA Inc. +1 503 615 3344 t +1 503 702 2067 m _______________________________________________ Tech mailing list [email protected] https://lists.lopsa.org/cgi-bin/mailman/listinfo/tech This list provided by the League of Professional System Administrators http://lopsa.org/
