Posted Nov 19, 2011 17:05 UTC (Sat) by jmorris42 (subscriber, #2203)
[Link]
> Do you also object to git's repository format then?
Nope. But note that a git repo that isn't cloned anywhere would be vulnerable to an attacker simply rewriting the hashes and thus being able to alter a repo. On a modern CPU hashing is fairly fast so any project that isn't as huge as the kernel could be compromised. That isn't a problem because anything important is replicated. That is the key to security, the hashing just improves it.
For syslog, replication alone is enough to solve the problem. Adding crypto foolishness and a bunch of binary fluff just makes it more complicated and reduces security. Put a log server somewhere on your network with only the syslog port open. If you are really paranoid you could store sha256 sums of each log as you rotate and pack it away on yet another machine or better on paper. Or just log the important entries on a line printer in real time as others have already suggested. Use a line printer without reverse paper feed and it is physically impossible to change the permanent record.
The Journal - a proposed syslog replacement
Posted Nov 19, 2011 22:32 UTC (Sat) by robinst (subscriber, #61173)
[Link]
I wasn't talking about security, just replying to this comment:
> So it replaces text files that can be read and processed with the standard UNIX tools with an undocumented binary format that can only be read by a single tool?
>
> Think I'll pass.
The Journal - a proposed syslog replacement
Posted Nov 20, 2011 1:56 UTC (Sun) by dlang (✭ supporter ✭, #313)
[Link]
the git repository format is well documented
The Journal - a proposed syslog replacement
Posted Nov 20, 2011 13:19 UTC (Sun) by robinst (subscriber, #61173)
[Link]
But was it well-documented from the very beginning?
The Journal - a proposed syslog replacement
Posted Nov 20, 2011 20:43 UTC (Sun) by dlang (✭ supporter ✭, #313)
[Link]
yes, the git on-disk format was well documented from the very first posts.