Quote:
Originally Posted by Ferus
Most people that have that amout of data, use their own cdn or someone elses, because downtime should never be a option.
I moved a single-instance tube less than a month ago, with more than 260 TB data, and no downtime whatsoever.
We used a Veeam proxy to start syncing to the second site 3 weeks in advance. It replicated the data every 12 hours the last week, to give us a understanding of how much data changed (for calculation freeze period)
- 5 hours before the move, we reset the Schedule to sync every 30 minutes
- 2 hours before the move, we initiated a freeze period for changes to scripts/appllications and uploads.
- Then the DNS switch
- Freeze for another hour after, to make sure all DNS had changed
- Shutdown old site
Make sure sitemaps are correctly updated and 404/301 redirects are set up correctly
|
Most of us with cdn go for the CDN cache, it still need to fetch from the original server. So if contents host on amerinoc, still unavailable, unless host else where.
About ur proxy syncing, was ur destination or receiving server with high traffic or just empty server doing the writing? Cuz when I use similar approach, from source to destination, but both servers have substantial amount of traffic every second, the I/O for the destination server or receiving server loaded super slow, it slow down the syncing speed. If I sync huge amount of contents, it took days. Last time I sync only 2 tbs of stuff, it took 6 days.