Hello all,We carried out an in-place upgrade on our production server on Saturday - going from 2008 R2 to 2014.We had tested this method out in dev/test and pre-production with only minor post issues to fix. However, on production we had an issue whereby checkdb was hitting 100% CPU and caused overnight processes to hang. The checkdb statement was terminated and disabled by a colleague at 1 am.Since then we have restored this database to a dev server and ran checkdb against it with no_infomsgs and all_errormsgs but it still hasn't finished since Monday morning!The database is just over 800 GB and whilst checkdb was crippling the cpu, logical reads are less than one. However, sp_whoisactive is showing that it has done 56 million reads so far and this number increases periodically so it looks like it keeps going back to re-check the database with a deep dive.Also, on a different environment, we ran check table statements and one of them took over 9 hours for a single table but came back clean (see attachment).Has anyone come across something like this before? I appreciate we need to wait for the output but the database is still in use in production and the mess will just get worse if it is indeed corrupted.
↧