Last night, I ran into all sorts of strange behavior on my laptop--unable to log in, normal system calls taking forever to complete, and all sorts of network trouble. This morning I backed up my home directory to my work computer's hard drive (discovering, in the process, that NTFS disallows all sorts of common and innocuous characters in filenames) and ran into several IO errors. Checked the hard drive and (despite SMART claiming everything was fine) it failed the read tests almost immediately. I managed to swap in a new drive and restore most of my files to a fresh copy of Ubuntu (with a customized version of tar to overlook the errors in the archive I made), but I still lost a fair bit of data.

Question: Why can't tar take an argument to skip over damaged sections of otherwise useable archives? A few IO errors at the beginning of the archive doesn't mean the remaining gigabytes of data are unrecoverable...

After a few busy hours, things are working smoothly again. Everything important for work has been recovered, and I'm polishing the asset tracking system I've been working on, in hopes of deploying it before my departure Friday. All things considered, I am quite thankful this misadventure occurred when it did--recovering data is the last thing I want to do in the few hours before leaving. :-)

Post a Comment

Comments are moderated. Links have nofollow. Seriously, spammers, give it a rest.

Please avoid writing anything here unless you're a computer. This is also a trap:

Supports Github-flavored Markdown, including [links](, *emphasis*, _underline_, `code`, and > blockquotes. Use ```clj on its own line to start an (e.g.) Clojure code block, and ``` to end the block.