Question: I’m trying to figure out what’s wrong with my brand new Seagate ST2000DM001(-9YN164) hard drive, if indeed anything. Irregularly, but seemingly every few minutes (every 3-7 minutes now in the time I’ve been timing it), it makes a short (0.2 seconds tops), fairly high-pitched, two-part sound. It’s definitely not normal “working noise”, and there doesn’t seem to be any correlation to disk activity.
The odd thing is that when I was using the drive with a USB-SATA adapter (one of those dongle things, not a plug-in cradle), sitting on top of the computer case with a paper magazine for insulation, it did not make any unusual noises during the copying of around 1 TB of data from the internal drive (which was showing signs of starting to fail, including reporting uncorrectable read errors during a routine file system scan). The change after that is that I removed the old internal drive, replaced it with the new one and re-fitted the SATA data and power cables.
I used GSmartControl, which presents smartctl data (I’m running Debian 6.0 on this system) to inspect the drive’s SMART data, paying particular attention to the Start/Stop Count (shown as: norm-ed value 100, worst 100, threshold 20, raw value 5), Seek Error Rate (66, 65, 30, 3981634) and Spin-up Retry Count (100, 100, 97, 0) (intermittent spin-up or seek errors was my first hypothesis, but the data doesn’t seem to support it). None of these values seem to be changing, and everything (both pre-failure and old-age) says failed “never”, which IMO is to be expected on a drive that has only been in service for a few days. The seek error rate raw value seems high but unless it’s growing rapidly I don’t see that as being a major cause for concern. Using the same tool I also ran conveyance and short self-tests, both of which completed without errors and show a lifetime hours value of 41 in the test log. I’m currently running an extended drive self-test (including a disk surface scan) but frankly don’t think that will turn up anything of much interest. One thing that certainly is interesting, however, is that in the a little over 15 minutes now that the extended self-test has been running, I haven’t heard any such noise, which, if nothing else, further points to something about the new drive being the culprit (I haven’t been doing anything else differently in this time).
Using hdparm, I disabled power savings by as root running hdparm -S 0 /dev/sda (where /dev/sda is the drive in question). This had no apparent effect.
Any ideas would be appreciated.
EDIT: Following the suggestion of @totaam, I turned the computer off completely (including the master switch at the back of the PSU), then restarted, and gradually added complexity. Everything was fine as far as I could tell through runlevel 1, but at runlevel 2 (which in Debian is multiuser with X, so basically the kitchen sink got thrown in at this point) the drive started exhibiting the same behavior again. So, I looked at what was started when entering runlevel 2, looking for any possible culprits, and found acpid as a clear candidate. Disabling that, however, doesn’t seem to have helped, even after re-running hdparm manually. I will take a close look at the BIOS settings as well, but in the meantime, here is a list of what was started entering runlevel 2, in order; does anyone else see any possible culprits? (And Seek Error Rate is still climbing, now up to raw value 4008400.)
- statd
- rsyslogd
- binfmt-support
- acpid (now removed)
- virtualbox-ose (host kernel modules)
- timidity
- dbus
- hald
- anacron
- atd
- gdm
- avahi-daemon
- cron
- cupsd
- kerneloops
- cpufreq
- ntpd
- postfix
- sshd
Answer: If this problem happens fairly regularly, I would take the time to narrow down its cause since this should not take you more than 20 to 30 mins of testing:
But from what you said, if it was me, I would just buy a new drive… my data and my time are worth more than one hard disk.