Re: Frequent Head Unload Problem

[Robert Hancock <hancockrwd@gmail.com>]

Gavin Cameron wrote:
> Hi Robert,
> 
> Thanks for the quick reply ... comments below.
> 
> Robert Hancock wrote:
>> Gavin Cameron wrote:
>>> Hi,
>>>
>>> It's been a while since I've been in touch. I previously had 2 problems:
>>> 1. PATA DVD writer not working
>>> 2. Frequent HDD head unloads
>>>
>>> I eventually caved in and replaced the PATA DVD drive with a SATA and 
>>> that works perfectly. However, this head unloading is really annoying 
>>> primarily because the noise the drives makes just doens't sound 
>>> healthy - it's a high pitch squeal and then a couple of clicks.
>>
>> Can you tell which drive is doing it? I'm assuming it's not both.. If 
>> you can figure out which one you can try some different APM values to 
>> see if that helps anything. You could also try disabling the standby 
>> timer in case that's doing it, with "hdparm -S 0".
>>
>> If none of those help, there may not be a lot else you can do about 
>> it. The OS really has no control over when the drive decides to 
>> unload, other than those settings..
> 
> I can't be sure which drive is making the noise - it may be hard to find 
> out as one drive is the root file system and the other is /home so both 
> are normally needed (unless I do a bit of hacking). The noise is the 
> same every time, so it may be just one of them. As it happens, /dev/sdb 
> (the drive allocated to /home) is the older of the two drives.
> 
> The unloading itself is very predominant within the first 15 minutes of 
> the PC being powered - happening between 1 and 3 times a minute. 
> However, eventually it decreases to perhaps once every half hour. It 
> often happens twice in succession too.

Can you try doing smartctl -a on both drives before and after you hear 
the noise and see if it seems to match up with any of the counters changing?

> 
> I've tried the following with no success:
> sudo hdparm -S 0 /dev/sda
> sudo hdparm -S 0 /dev/sdb
> sudo /sbin/hdparm -B 253 /dev/sda
> sudo /sbin/hdparm -B 253 /dev/sdb
> sudo /sbin/hdparm -B 128 /dev/sda
> sudo /sbin/hdparm -B 128 /dev/sdb
> ( as well as 254 and 255)
> 
> I know you're saying that there's not much else can be done from an OS 
> point of view, however ... this particular PC didn't make this noise 
> until kernel 2.6.19 (I think it was) that started using libata. Also, 
> it's a dual boot machine (although I hate admitting that to people - but 
> still need it for gadget firmware updates) ... and when it runs XP, the 
> drives never make a noise. Not that I'm trying to say that Windows is 
> any good for anything ... it's just an observation.
> 
> There was also the previous problem with the PATA DVD drive not working. 
> So I still believe that the libata kernel module has some problems with 
> this chip set (perhaps combined with particular drive usage, I'm not sure).
> 
> Hopefully, there may be some other file or command output that I can 
> provide to you that will help to diagnose this. If so, please let me 
> know and I'll gladly provide it :-)
> 
> Best regards,
> Gavin.
> 
>>
>>>
>>> Anyway, a reminder of the basic system:
>>>     * ASUS M2V M/B (VIA K8T890 / VIA VT8237A chipset)
>>>     * AMD Athlon 64 X2 6000+
>>>     * SATA LiteOn DVD RAM/RW
>>>     * Ubuntu 9.04 (Beta) - kernel 2.6.28-11-generic
>>>
>>> Since I last was in touch, I've converted the system to Ubuntu (I was 
>>> hoping that it was a Fedora quirk) with no change in head unloads 
>>> with either 8.10 or 9.04.
>>>
>>> I've tried the following in /etc/rc.local :
>>> /sbin/hdparm -B 254 /dev/sda
>>> /sbin/hdparm -B 254 /dev/sdb
>>>
>>> Which makes no difference at all. Also, I tried setting 255, but that 
>>> also had no affect.
>>>
>>> I've included all of the normal files that I think you may need:
>>>
>>> sudo dmidecode > dmidecode.txt
>>> sudo hdparm -I /dev/sda > hdparm-sda.txt
>>> sudo hdparm -I /dev/sdb > hdparm-sdb.txt
>>> sudo smartctl -a /dev/sda > smartctl-sda.txt
>>> sudo smartctl -a /dev/sdb > smartctl-sdb.txt
>>> lsmod > lsmod.txt
>>>
>>> Hopefully you can find the root of the problem.
>>>
>>> Best regards,
>>> Gavin.
>>>
>>
>>
>