My migration to a new server for all the web-based projects I’m hosting has been going great … until now. This morning, I woke up and was working on my main computer when I noticed that my server wasn’t kicking up every few minutes to refresh the OSM planet database. I turned on the screen and was staring at a failed update due to an I/O block error. I unmounted the offending hard drive and ran fsck on it – leading to the photo below.
Fsck showed tons and tons of bad blocks. I eventually killed the process after using my phone to mash down the “y” key (this, by the way, is why they have the “-y” option on fsck.
In any case, it appears to only happen on writes (as far as the nominatim update goes), and I can still query the database. But I’m sure this is a sign that the “new-ish” (3 months old) hdd is failing already. I’m going to write back to the dealer who sold me the system, but I’m thinking I’m unlikely to get a refund, plus I have a working full planet install of nominatim, so I’m not sure if it is worth the hassle.