Part of my migration strategy of moving to Amazon EC2 for our servers involves making use of symlinks to keep installs and files in their 'standard' locations on the servers but the actual storage of logs files, data, etc. on EBS storage, for persistence. After the server starts up, I run scripts that create symlinks to the config files and data stored on the EBS to 'convert' the server to the setup I need.
Since I'm not a true Linux sysadmin (small company developer), I'm nervous about any sort of gotchas I might be unaware of by using symlinks. Things like breaking software packages or other difficulties where the applications might not like utilizing symlinks are what I'm concerned about.
Are there any common gotchas using symlinks or are they pretty foolproof?