Hard drive constantly accessed with SuSE 9.3
I just installed SuSE 9.3 after working with 9.2 for a while and have noticed that my hard drive light never goes off, and (over the loud fans) I think I can hear it being accessed too (hence the light),
Now with SuSE 9.2 I distinctly noticed that light would remain off unles I was actually doing something... So, I'm not exactly sure what to do... I tried booting with acpi=off, but it didn't really do anything I've got: two Seagate 200gb SATA drives SuSE 9.3 Kernel 2.6.11.4-20a-smp Intel 915pbl motherboard I just don't want my hard drives to be constantly accessed... less performance, less life... |
"I'm not exactly sure what to do."
Run the top command. It will show you what is running. That will at least give you a list of suspects to examine. --------------------- Steve Stites |
Unfortunately, that's not yeilding any results. I've taken it down to runlevel one and still tried to kill even more processes, but even then, none seem to be doing anything out of the ordinary... No unruly cpu usage, or anythig just crazy...
I seem to recall it also being this way when I put in the first SuSE 9.3 install CD, I'm starting the think it may be the Kernel... but I'm not really experienced enough to really know what to investigate the whole kernel area (bad terminology) |
Is the HDD activity light blinking or solid? If it is solid (not blinking) then hardware is suspect; pull all HDD cables and then re-seat them one by one, and make sure the BIOS enumerates each device properly.
|
The light is solid, but why would it change from blinking to solid between SuSE versions? I've gone back and forth a couple times now and have the same results:
Nice blinky blinky in SuSE 9.2 Solid Lighted Badness on SuSE 9.3 |
Perhaps a cable worked it's way loose?
A drive died? Sunspots? Gremlins? ;) |
might the contast scanning by due to beagle?
|
What are your swap parameters for 9.2 vs. 9.3 (on/off)? Also have a look at /etc/fstab and compare the swap-spaces...
|
JZL240I-U>
When a computer is hitting swap, the drive activity light would be blinking, not steady/solid. Solid lit means that a request has been submitted and is pending, not completing; usually indicative of nonresponsive hardware. Oops, and my humblest apologies for the atrocious error above where I typed "it's" when I should have typed "its." Darn. |
Well, first off, I kept my swap partitions the same. It's about 1gb on sda, while I've also got my root on sda. I've got sdb reserved for /home. When I start up any of the little monitors that come with X to look at I/O stuff (like hdds, eth0, cpu) and it seems like it's hitting sdb, but not for very much...
Anyway, I'll also run some tests on the drive, see if there are loose cables and stuff as soon as I get home from work... It just kind of floors me, cause I went from 9.3->9.2 back to 9.3 and while I was using 9.2 it was fine... Darn sunspots... |
You know, there could be another reason that hardware is not responding to requests properly: your kernel may not be loading a 100% compatible module for the drive controller chipset. Try recompiling the kernel and make sure you include support for the exact chipset in your system. Personally, I'd check the cables, etc. first.
|
Constant Drive Access Maybe is Not Hard Ware Problem But A very Active Cron Daily
Before getting hung on Hardware Issues Check.
The Schedule for Your Cron or Update schedule for RPM look up and Scan Intervals. The system when set-up default has a schedule that pretty muchs runs all the time. Also a feature of the Cron as set-up is it runs Silent. That converts to never seeing direct program access in task list. The first thing to check the Messages log and look for actions of Cron. Login as root and Modify The schedule down to something other than run taks at about every 45 minutes, which was the default for both suse 9.2 and 9.3. Access to CRON is under system tools when logged as root. The cron schedule is not part of system menu tools when logged in as normal user. The only schedule control is ability to create user and not change root schedules. To heck for messages related to the problems use the command below. While logged in as root Enter the following. tail -400 - f /var/log/messages for the above the number is lines to look at in the log. This will give the last actions for the current session. To get out of the listing mode press the Control and C. If you think some kind of hardware problem exists check the warn log. The log could show disk access errors. Linux does monitor and report such things. The log is accesed as root tail -400 -f /var/log/warn The other action you can do if hardware is an issue is start the SMART Tools from root The action is /etc/init.d/smartd start You may have to edit the smart.conf file to indicate which drive or drives to monitor. From the command prompt you can manually get a drive status and error list at prompt /etc/init.d/smartd status I personally monitor drive health using the WEBMIN remote tool. I then can check status for the three linux boxes on out network with out have to hand check all the time. Hope that you find it is software and not hardware as a problem. ;) |
Thanks for the info! I've activated the smartd and I'm also going to check out the hardware as well.
You know, you've just got to love webmin! It comes in handy SO many times. As far as recompiling the Kernel goes, is there a certain site you'd recommend for a howto/reference/help. That's something I've never really taken part in.... It always makes me a little nervous to think of recompiling a kernel... |
Quote:
|
Quote:
|
All times are GMT -5. The time now is 09:24 AM. |