So I've written before on my use of Spideroak. The service has proved to be a fantastic help and runs nicely in the background but I've hit a few snags using the program itself.
Going in to change the details of what’s being saved to the Spideroak servers, I find that the client gives me a lot of the “spinning beachball of death”
Ok, I don’t have to change the saved folders very often but like today where I decided to remove a .git folder from the backups (and from being backed up), the client felt like treacle. It took me at least 5 minutes to accomplish a task that should have taken a few second (Selective sync on Dropbox? Can select my folder I don’t want synced in, as quickly as it takes me to find the folder in the menu!) – Why cant Spideroak do the same? At the minute, I browse to the folder, select unsync and then wait for a bit as the beachball spins. Then it slowly moves it on the server to the deleted items – which is fair but why does that slow down my local machine?
It could have been the fact that I have Spideroak to automatically send data to the servers as soon as it’s changed on my machine. Which leads to my next niggle.
Lion Auto Save
Lion auto save is a really nice feature – I really like it (on most programs – I don’t like it on programs where I want to experiment with data and potentially need to roll back. Maybe I'm just not used to, or trusting enough of the Lion versions). Regardless, I think my above data backup schedule is leading to potential issues with Spideroak.
As I mention above, I have it set to automatically send as soon as the file has changed. This has led to a few massive backups (and if it wasn’t for Spideroak’s de-duplication feature, I might have run into space trouble!) of a few programs data – specifically Trailrunner which had a 3.5GB backup size on the Spideroak server and only 43MB on the disk because it saves the file extremely often (every change to a GPS track seems to set it off – Dropbox and Spideroak must hate me/it!).
I think a less stringent backup regime (maybe scan it every half hour or every hour like the Time Machine backup) might help reduce the beachball of death (perhaps) and prevent what appears to be massive amounts of backed up data from the Lion autosave function.
It would be interesting to see how much network traffic/bandwidth and load the addition of autosave has placed on services such as this – I doubt it will have added to much to the load because at the end of the day, Mac market share isn’t fantastic and then the uptake of Lion (and Lion autosave ready apps) will also affect the upload bandwidth.
In the mean time, I have some clearing out of Spideroak to do to ensure it only saves versions of files that ideally I want!