bkil vs. programming

Here is a link where you could theoretically monitor Google outages: https://www.google.com/appsstatus#hl=en&v=status I had a problem with Google services just recently. However, as I wanted to verify whether the problem is on my side, this monitoring site was down as well. At the same time, the third party reports on alternative sites started to grow: http://outage.report/google http://downdetector.com/status/google http://aussieoutages.com/status/google This highlights why we should not put all our eggs into the same basket.

While it may sound like utter nonsense at first, on second thought maybe it really is: cron.standard: please provide configuration facility to avoid given filesystems Someone should really be thinking more about a serious reminder system. I've put together an iotop oneliner that logs disk actions with timestamps which pointed me to anacron. Daily mlocate updating was easy to reproduce and fine tune: simply add your HDD mount points to tho PRUNEPATHS variable of /etc/updatedb.conf . On the other hand, it took a few more attempts and many days in total to track down what causes the HDDs to still start up. When trying to further examine the issue by hand, /etc/cron.daily/standard didn't start them up on first look, perhaps because of caching. However, today after randomly testing things again, it finally did, which took me to adding CHECK_LOSTFOUND=no to /etc/default/cron . It's a pity though, that I'll have to use an alarm clock from now on.