![]() I mean, an example is if their laptop is stolen or died 5 minutes before the most important presentation they have ever given. Backblaze is a 24/7/365 operation, customers NEED and EXPECT to be able to access that one particular file that is critical to them within seconds. Backblaze does about 265 restores for customers EACH DAY on each cluster authority. That is about 105,000 customers, but it would depend on which cluster authority has problems. The reason the Backblaze server engineers need to be called on the weekend to fix this is that if there are 7 cluster authorities in the world, at least 1/7th (14%) of Backblaze customers are in your situation and unable to sign into their accounts which then prevents them from accessing their data. A totally different issue is a network card goes bad in a cluster authority, and the datacenter techs need to run over, swap a new card in, and bring it back online. All the data is perfectly safe in this scenario, but you cannot login. So the mapping server could not reach the cluster authority. But here are some things that have gone wrong in the past: an international network link (think undersea cable) had an issue, like a router died or a ship anchor cut it, doesn't matter. If I still had access to servers, I could tell you what actually went wrong. One is in Amsterdam, one is on the East Coast, one is in Phoenix Arizona, etc. ![]() The important part here is those cluster authorities are ALL OVER THE WORLD. There are about 6 or 7 of these cluster authorities in the world, so as you login Backblaze needs to "lookup" which one your account is on to send you over there. This is one way Backblaze scales: your account is bound to that cluster authority for life. You aren't getting to that point yet because the "cluster" you are on has this one computer called a "Cluster Authority" contains your login credentials, not any of your actual data. Only the first: "this should not occur and it is fixable".Īll of your data is stored across thousands of "pods" organized into "vaults", which have literally nothing to do with this particular issue (logins). "it shouldn't happen but it's fixable" or " there's a risk I might not get my data" Either way it's a pretty big deal and Backblaze needs to fix it in the next few hours. That might be a network problem, or it might be your cluster authority is offline, or some other issue. What the Backblaze problem that is occurring is that your particular "cluster authority" cannot communicate with the global mapping server that maps your email address to your cluster. The critical information for support is this line: You can open that file with WordPad on Windows and TextEdit on the mac. On Macintosh: /Library/Backblaze.bzpkg/bzinstall.xml On Windows: C:\Program Files (x86)\Backblaze\bzinstall.xml In the very first support case, attach this file from your local computer: You need to open a support ticket by going to and scrolling all the way to the bottom. There was an issue fetching your account, please try again later. ![]()
0 Comments
Leave a Reply. |