Recommended Posts

Just found out that RAID arrays can have alignment issues, separate from the well known 4K sector issue.

I suspect it to be the cause of slow transfer rates on my Adaptec RAID 3805 RAID6 (5x Hitachi 2TB 7200 RPM) array.

Nodesoft Disk Bench only gets 100MB/s writes and 80MB/s reads (these figures are more in line my Windows File Copy performance)

http://www.nodesoft.com/diskbench

But I don't know why the following benchmarks report so much higher

Atto/HDtach:

ATTOANDHDTachbenchwithRAID63805Hitachi2TB.jpg

Crystal:

CrystalDiskbenchwithRAID63805Hitachi2TB.jpg

However,

I think I have confirmed that my array is not aligned, but am not sure.

State of Array:

aligninfo1.jpg

Results:

Alignment Calculator:

http://www.techpowerup.com/articles/other/157

aligncalc.jpg

Alignment Checking Script:

http://www.ocztechnologyforum.com/forum/showthread.php?57599-RAID-Non-RAID-Partition-Alignment-for-max-performance&p=518021&viewfull=1#post518021

  Quote
Device ID: Disk #2, Partition #0

=======================================================

Total Volume Size: 5 TB

Starting Offset: 134235136 (131089 KB)

Partition Alignment: ** NOT ALIGNED!! **

VM Check Alignment Tool:

http://ctxadmtools.musumeci.com.ar/VMCheckAlign/VMCheckAlignment10.html

vmcheckalign.jpg

My Calculation:

Formula: ((Partition offset) * (Disk sector size)) / (Stripe unit size)

Source: http://support.microsoft.com/kb/929491

Partition off set = 134235136b

Disk sector size = 512b

stripe unit size = 262144b

block size = 512b

answer = 262178

This is an integer, thus indicates the array is aligned. But this conflicts with the conclusions of the other calculators, so WTF? I have no idea what I'm doing lol :weep:

Array Info:

Adaptec RAID 3805 RAID6

HD 5x (Hitachi 2TB) Sector Size: 512Bytes

Stripe Size: 256K

Disk Format: GPT

Windows Disc mode, Basic or Dynamic: Basic

This topic is now closed to further replies.
  • Posts

    • Missing the Streets of Rage 2 Classic, the link provided its for Streets of Rage Classic both are free
    • Mastodon updates terms of service to ban AI model training on user data by David Uzondu Mastodon has updated its terms of service to explicitly prohibit scraping its platform to train artificial intelligence models. The new rules, which kick in on July 1, make it perfectly clear that using automated tools to slurp up user data from its main server, Mastodon.social, for LLM training is a big no-no. Neowin received a copy of an email sent to users, notifying them of the change, which introduces new language prohibiting the "scraping of user data for unauthorized purposes, e.g., archival or large language model (LLM) training." Here's a snippet from the updated terms of service: This policy change comes at a time when users are getting increasingly ###### off about their public posts becoming free fuel for the AI gold rush. In fact, this is probably good news for the same crowd over on Bluesky that freaked out after a massive, user-traceable dataset of their public posts was compiled and uploaded for AI research. AI bot scraping has become a huge problem for everyone from giants like Reddit, which is now suing Anthropic, makers of Claude, for training on its posts without a license, to even Neowin readers, like Gerowen, who noted how a swarm of bots, including one Claudebot (you don't say!), hammered his personal server with over 700,000 requests in 24 hours, putting a huge strain on his "home NAS running on an old PC tower in the back woods of Kentucky." It is important to remember that Mastodon is a federated network. These new terms apply specifically to the Mastodon.social server, which is operated directly by Mastodon gGmbH. This means that while users on the main instance are now protected, those on other independent servers in the "fediverse" will only get the same protection if their instance administrators adopt similar terms. The company is globally enforcing a new minimum age requirement of 16 for all users, raising it from the previous limit of 13.
    • Keep in mind that updates for it end on Oct 13, 2026. While this may not matter much for those who don't care about features, it might for fixes, and it certainly would for security.
  • Recent Achievements

    • Explorer
      JaviAl went up a rank
      Explorer
    • Reacting Well
      Cole Multipass earned a badge
      Reacting Well
    • Reacting Well
      JLP earned a badge
      Reacting Well
    • Week One Done
      Rhydderch earned a badge
      Week One Done
    • Experienced
      dismuter went up a rank
      Experienced
  • Popular Contributors

    1. 1
      +primortal
      699
    2. 2
      ATLien_0
      275
    3. 3
      Michael Scrip
      219
    4. 4
      +FloatingFatMan
      190
    5. 5
      Steven P.
      146
  • Tell a friend

    Love Neowin? Tell a friend!