I got the following emailed to "root" by "smartd": From: root <root@diet-anarchy> To: root@diet-anarchy Subject: SMART error (CurrentPendingSector) detected on host: diet-anarchy Date: Mon, 20 Aug 2007 18:33:56 -0400 This email was generated by the smartd daemon running on: host name: diet-anarchy DNS domain: [Unknown] NIS domain: (none) The following warning/error was logged by the smartd daemon: Device: /dev/hda, 1 Currently unreadable (pending) sectors For details see host's SYSLOG (default: /var/log/messages). You can also use the smartctl utility for further investigation. No additional email messages about this problem will be sent. ** I got the following emailed to "root" by "logwatch": --------------------- Smartd Begin ------------------------ Currently unreadable (pending) sectors detected: /dev/hda - 11 Time(s) 1 unreadable sectors detected Warnings: Warning via mail to root: successful - 1 Time(s) Sending warning via mail to root ... - 1 Time(s) ---------------------- Smartd End ------------------------- *** There are two usability problems here. 1.) The wording of these warning messages is rather technical. Most people - who are not experienced command-line Linux users - would not know what to do next. 2.) These messages are being e-mailed to "root". One would need to either actively check root's email, configure sendmail to send root's email somewhere else, to even receive these messages in the first place. Anaconda does not prompt users to do either of these things, and I seriously doubt that most everyday users would actually do them. These are important messages, which after some amount of research I decided meant that my hard drive was beginning to fail, and that it was time to replace it with a new one. This is important enough that not receiving or understanding these messages seems unsatisfactory. One possible solution is to provide a pop-up warning, similar to those generated by SELinux or Network Manager, to any users logged into GNOME. Better online documentation to help novice users diagnose their hardware problem could be helpful since there are a variety of messages which might be generated. There might also be completely different solutions which are also satisfactory.
The action that smartd performs on error is configurable and the configuration is described in the documentation even with examples. You can configure smartd to use notify-send, but since smartmontools package is more important and useful on servers (it's going to be turned off by default in F-8) I have to assume there will be no user logged into GNOME. Then the only action that makes sense is to send an email to root. The manpages are browsable in yelp and I consider them to be really well written. Much better than I could do with my English...
For users that don't even know enough to check root's email, it's not a viable solution to expect them to re-configure smartd to take some different action upon discovering a problem. I am running Fedora 7 on a laptop, and I was very happy that smartd was enabled by default. If it hadn't been, I wouldn't have known to replace my hard drive. I don't see why it's any less useful for desktop users to get notified about imminent hardware failure; they are even less likely to be making regular backups of important data. I would definitely recommend enabling smartd by default in Fedora 8. What would be the downside?
Having smartd turned off makes sense assuming users are clever enough to know what they want and since all of them want fast system, nothing that's not required should not be started by default. It's the user who decides what to run on his machine, not the packager. And I think this is a good approach.
The point of enabling smartd by default is to do the right thing for users who *aren't* clever enough to know that smartd exists in the first place, much less figure out how to turn it on and receive its messages. I'm sure everyone wants a fast system, but it seems a bit more important to know whether or not they are about to lose all of their data due to hard drive failure.