[clug-talk] Thorough disk check/stress test utilities?

Curtis Sloan curtis.sloan at shaw.ca
Thu Jan 6 13:28:02 PST 2005


I have been experiencing some odd behaviours lately in an otherwise very 
stable desktop box (mostly hangs, some odd clunks now and again).

At first I thought it was RAM since I tried having an upgrade done in rather 
dry, cold weather.  However, I ran memtest86 for a few hours afterwords with 
no problems.

The clunking I thought was a disk spinning up from power saving mode (since I 
semi-recently added -S240 to hdparm settings).  But lately they seem to be 
happening inside 20 min. of last disk activity.

Also, I did do a standard fsck after a couple nasty hangs in row (for both 
ext3 and reiserfs) and reiserfsck found a few corruptions and fixed them.  
Naturally, I didn't bother to note which disk or partition, since I still 
thought at that point it was a byproduct of a RAM-induced hang.  :-P

So I'm starting to strongly suspect one particular drive (I have 3 x 10GB IDE 
drives).  I'm backing up all my data right now.

I've got 1 x FAT32 shared partition, 2 x reiserfs partitions and 2 x swap 
partitions.  (Yes, 2 swap... I had extra space to kill and didn't need a tiny 
128MB partition for data.  Swap priorities are set accordingly :-P).

I want my system back to its stable old self.  If that means losing 1/3 of my 
storage, then so be it.  I just don't want to do it on a hunch.  So after 
backing up my data, what can I do to:

1)  Thoroughly check all file systems, including swap.
-note:  I tried a thorough GUI Scandisk from Win98 which resulted in a hang.

2)  Stress test the disk to see if I can induce/reproduce a problem?
-note:  the disk activity LED stays solid red on a hang w/ no disk activity.

Sorry for the verbose post.  I'm a little caught up in it at the moment.  It 
seems I haven't done backups for awhile, so I've been a little nervous.  :-P

Thanks in advance,
Curtis S.



More information about the clug-talk mailing list