Take the 2-minute tour ×
Server Fault is a question and answer site for professional system and network administrators. It's 100% free, no registration required.

This question already has an answer here:

I recently set up a server running Ubuntu 10.04 LTS on a Dell PowerEdge 2950 server - it has 6 500 gb 7200RPM SATA drives setup in a RAID 10 config.

I am seeing extremely poor disk performance - the RAID array reports all disks are normal and using MegaCLI, it looks like the BBU is fine.

hdparm -tT /dev/sda reports:

  • Timing cached reads: 90 MB in 2.05 seconds = 43.96 MB/sec
  • Timing buffered disk reads: 24 MB in 3.11 seconds = 7.72 MB/sec

So as you can see, it takes forever to something as simple as an apt-get upgrade and even logging into the server.

How do I go about troubleshooting what is causing this? I upgraded the firmware on the PERC 6i RAID controller to the latest, but didn't see any improvements.

share|improve this question

marked as duplicate by Chris S Jul 12 '13 at 14:00

This question has been asked before and already has an answer. If those answers do not fully address your question, please ask a new question.

    
How old is the card? I have come accross one or two bad memory chips that put a severe damper on performance, even kept the OS from loading. Also, does this particular model come with a cache battery? –  hydroparadise Mar 30 '12 at 20:16
    
yes, it comes with a battery and using the MegaCLI tool provided by LSI, it says everything with the battery is a-ok. From the logs, it doesn't appear to be forcing it into write-through mode - drives are set to write-back but nothing is set for read-ahead. The card is 4 years old. –  vocoder Mar 30 '12 at 20:46
    
I know with IBM systems, they actually have a specified number of days that the card is allowed to use for a given battery. Even if the battery technically still holds a charge, the system acts if the battery is dead once it's past that period of time. Also, the RAID only performs as fast as the least performing HD. Could be an individual HD. –  hydroparadise Mar 30 '12 at 20:59
    
I believe the Dell PERC's do a battery check every so often so if it reports ok, battery should be ok. However hydroparadise raises a good point about checking the drives - there could be one slow drive holding up the others. Have you ever rebuilt the RAID? I've seen an older PERC "successfully" rebuild a RAID then have terrible performance afterwards. Why not enable read ahead cache - this is what will help on reads and you have mentioned slow reads. –  Robin Gill Mar 31 '12 at 15:52
add comment

3 Answers

Recently we had a problem on one of our Dell DAS units and enabling Write Caching policy at controller level to "Forced Write Back" tremendously improved the performance.

here is the setting

http://support.dell.com/support/edocs/storage/Storlink/H700H800/en/UG/HTML/chapterb.htm#wp1077064

I hope this helps.

share|improve this answer
add comment

The same problem but with PERC 5/i was fixed with "Forced Write Back".

share|improve this answer
add comment

I had similar issues with PE2900 and PERC 6/i, very slow writes (~ 40 MB/s) with Ubuntu 12.04 and I noticed that when I boot from the CD in rescue mode the speeds were good (~ 200 MB/s), so I started blacklisting linux modules and found out the cause - i5000_edac and edac_core.

vi /etc/modprobe.d/blacklist-edac.conf

blacklist i5000_edac
blacklist edac_core
share|improve this answer
add comment

Not the answer you're looking for? Browse other questions tagged or ask your own question.