Red Hat Bugzilla – Bug 98376
smartd misses /dev/hdm
Last modified: 2015-01-04 17:02:35 EST
From Bugzilla Helper:
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.4) Gecko/20030619
Description of problem:
smartd fills the syslog with errors about being unable to open() SCSI devices up
to /dev/sdz, but it stops IDE scanning at /dev/hdl, missing two of my hard disks
named /dev/hdm and /dev/hdo. Any chance of extending smartd such that it goes
up to /dev/hdt, which seems to be the last hd device in /dev, and of silencing
the errors that needlessly fill not only the syslog, but also the LogWatch summary?
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1.Get a machine with 8 IDE channels and install a disk in the last such channel
3.Look at the syslog
Actual Results: It didn't even try to look for /dev/hdo or /dev/hdm
Expected Results: It should go up to /dev/hdt.
(1) You can stop smartd from scanning at all, by creating /etc/smartd.conf.
Please see man smartd and man smartd.conf
(2) I did't realize that IDE devices could extend beyone 12 (/dev/hda-/dev/hdl).
If no config file is given, we can trivially modify it to scan beyone /dev/hdl.
How many additional devices are needed?
(3) If you want to only scan ATA and not SCSI, you can use
DEVICESCAN -d ata
(4) I'm not sure what version of smartmontools is currently in Fedora. I
suggest upgrading to 5.20 or 5.21. See http://smartmontools.sf.net
IDE can go upto 20 disks...
OK, I'll fix it in smartmontools release 5.22. We'll scan 20 devices /dev/hda-t.
Thanks, I have set up a smartd.conf a long time ago to work around the bug, but
I still try DEVICESCAN every now and then to see whether the default scanning
doesn't cover all devices it should. Thanks for fixing this!
Somewhat off topic: I do a search query of bugzilla for smartctl,smartd,
smartmontools,smartsuite every couple of days. This is the first time that I
got your three-month old bug report in the response. Any idea why? Was
bugzilla just updated and a search bug got fixed?
It was originally filed against a private alpha release. I only changed it to
Fedora Core and opened it up today.
This has now been fixed, and will be in the upcoming smartmontools 5.22 release.
If you would like to test it, please download the code from the SourceForge CVS
pserver - instructions are at http://smartmontools.sf.net/
Erhm... kernel-utils-2.4-9.1.101.fedora was updated to 5.21, so it still
doesn't have the fix :-(
This is fixed in the smartmontools 5.22 release. I am sure it will migrate to
Fedora reasonably soon.
It's fixed in rawhide, and probably has been for quite a while. Thanks,