Jump to content

Slow parity check since upgrading to rc4 (and rc5)

kgsb

By kgsb
in General Support

Recommended Posts

kgsb Noob

kgsb

Posted
Posted

Prior to RC3 and now RC4 the parity check speed was between 40 - 60 MB/sec. After updating to RC4, started a parity check, speed now stays about 8 MB/sec. Cancelled rebooted and restarted many times, no change. Loaded RC5 today, no change.

Can someone take a look at the attached log file and see if there is a problem. I have not yet tried loading an earlier release to see if the problem goes away, if there is nothing obvious in the log, I will try that next. 

syslog.txt

Link to comment
RobJ Mentor

RobJ

Posted
Posted

Syslog looks OK, no glaring issues, no drive issues at all.  Attach a diagnostics dump, so we can check the SMART reports.

 

Sometimes, the first indication of a drive issue is unusual slowness in I/O to a drive.  Internally, it retries and retries when one or more sectors are getting hard to read.  Only when it has to give up does it report the issue.

 

You have one issue that could be changed, that may be affecting performance a little.  Your 4 motherboard ports are still set to IDE emulation.  In the BIOS settings, look for the SATA mode for them, and change it to a native SATA mode (preferably AHCI but your board may be too old), anything but IDE emulation.  It should be a little faster, and a little safer.  For kicks, you may want to test your drive speeds, before and after.  The command hdparm -tT /dev/sdX provides the top speed of the drive (use the very last number, ignore the rest).  It's usually best to run it 4 or 5 times in a row, throw out any far off ones, and average the rest.  Change sdX to the correct drive symbol for each data and parity drive (eg. sdc, sdg, sdi, etc).

 

Another thing that can significantly impact parity speed is refreshing the Main screen while it is running.  Current versions turn off Main screen automatic updates, and show a Refresh button, which should not be pushed very often.

Link to comment
enetec Apprentice

enetec

Posted
Posted

I unzipped the smart folder from the diagnostics file above, see attached. It is missing the flash drive file which only had a couple of lines.

I can run another diagnostics and upload again if required.

 

You have some pretty old drives and two of them are showing some issues...  :-\

 

My tips:

 

- change SATA cable and, if possible, SATA port, of your Maxtor HDD and then test its surface with a HDD Regenerator live cd. This drive has been logged for UDMA errors on both SMART and unRAID and seemed to become unresponsible since of bad connections or internal issues. Data wil not be touched during process;

 

- make same surface test to your Seagate with serial ending with C4T8. It has already marked some bad sectors and expect to find more and/or a lot of delays (which aren't god too...);

 

I would test them in two pass for safety reason: first one do a 'Scan only without repair'. If the disk tend to disconnect or hang during process immediately backup all its contens as fast as you can and substitute it with a new drive in array.

 

If they end this first pass (even with delays or bad sectors found...), go for a second pass with repair enabled this time.

 

Eventually post screenshots of stats result at the end before re-adding them to array.

Link to comment
kgsb Noob

kgsb

Posted
  • Author
Posted

Thanks for the suggestions, will do over the weekend. I have a couple of 3 TB replacement drives but I have been waiting to get V6 stable. The plan is to convert to XFS during the process. This is an old system, but it has been extremely reliable over the years, it is starting to show its age with the move to V6.

Link to comment

Archived

This topic is now archived and is closed to further replies.

Go to topic listing
×
×
  • Create New...