Why does SCSI Hard disk show a slow transfer rate

Posted on 2010-09-15
Medium Priority
Last Modified: 2012-05-10
I have three Seagate SCSI hard disks attached to Adaptec 39160 controller card. I  have Asus P6T V2  motherboard, Core i7 processor and have triple boot seqence of Windows XP, Vista and Seven. Previously the hard disks attached to this adaptor were showing a transfer rate of 160 mbs the maximum for this card. But at times they showed less then that like 40, 80, 160 mbs for each hard disk. Then what I used to do was I pressed the power leads to the hard diosks because the power leads had become loose. Since this became a problem I changed the Power Supply from Asus 460 volts to Corsair 850- volts. That solved the problem for some days. But then again I am having problems. I suspected the SCSI cable. I took it off and reinstalled now all the SCSI HDDS show 40 mbs at start up instead of 160 mbs. It is a old SCSI cable but apparently the pins look stable. The ones which were bent were straightened out. What could be the reason the SCSI cable which may have undected bent pins or the SCSI Adaptec 39160 Adaptor
Question by:SALEM586
  • 2
LVL 47

Expert Comment

ID: 33685451
First SCSI (parallel SCSI) somewhat uses lowest common denominator if all disks are on same bus.  It is arbitrated, so only one thing goes on at a time.  Send some data at 40Mbs, and everything waits that much longer.  Any disk with crappy connector, or that is throwing certain types of errors, can cause the initiator to crank them all down.

yes check cables.   But looking at pins isn't good enough.  Unless you have about $5000 worth of equipment, best you can do is buy some software that will really dig into each disk and report link speeds over time, and tell you what is going on.  Not even a decent digital oscilloscope will tell you what you need, other than voltage levels. Certainly a voltmeter isn't going to be very useful.

Even a new cable won't fix problem if a certain drive is root cause.  Not only that, but disk drives can potentially be programmed to only run at certain speeds.  


Author Comment

ID: 33688507
I shall be obliged if you could further explain-

      Crappy connector means the the SCSI connector to the hard disk not the the power connector. This SCSI cable has five connectors. So if one connector is faulty it will bring the transfer rate of all the the SCSI hard disks. The three SCSI Hard disks are on the same bus. This SCSI adaptor supports two buses. If I put any SCSI hard disk on the second bus on another SCSI cable would that make any difference and will the the SCSI hard disks boot as they did before. When the transfer rate goes down to 40mbs it is very difficult to work in windows because all the processes become so slow.

 The pins on the SCSI cable gets to get bend and deformed after all there are 68  pins. They are not easily attached to the hard disks unlike SATA cables.Indeed I would like to upgrade to SAS which is point to point serial SCSI interface and having SAS cables which have much less pins and so are more robust and stable. But the cost of a SAS adaptor and SAS hard disks deters me.

The three SCSI hard disks are two SeaGate ST373455LW running at 15000 rpm. The third is Seagate ST314680LW which is running at 10,000 rpm. You also mean to suggest that in addition to a faulty SCSI cable that any error in in any of the hard disks can bring down the transfer rate. Is that correct?
So the faut lies either with the SCSI Cable or the hard disks not with the SCSI Adaptor.

Just of academic interest and since you are very qualified to answer since this SCSI card is installed on a PCI slot, its inherent transfer rate will not exceed 66 mbs of the PCI slot. To get better transfer rate either the PCI-X slot or the PCI-E slot should be used.
LVL 47

Accepted Solution

David earned 2000 total points
ID: 33688666
Well, it is more complicated than that.   A bad something can sometimes create noise and lousy signals, these translate into parity and CRC errors.   The adapter learns that speed at u160 isn't reliable, so it drops things down to minimize and/or eliminate the errors.   A disk drive can be root cause. Some hardware/driver combinations won't crank speed up either.   RPMs for disk drives have nothing to do with it.

PCI slot is not 66Mbs, it is 66, 133, 266, 33, depending on flavor of PCI and/or PCI-X.  This is bus rate, how fast the data gets to your CPU, not anything to do with the SCSI interface. If you saturate the bus, then speed will still stay at U160  or U320.  It is just that the queue will be full so the disks and/or initiator (controller) tells each other to stop to catch up.  Other things into the mix are priority levels.   It is important to get enough bandwidth on backplane.  PCI/PCI-X also shares bandwidth.  Every card in the bus shares as well.  So in grand sense of things, if you have 2 x 133mhz PCI-X slots, and put in 2x133Mhz PCI-X cards, then effectively each card will get 66Mhz.  

That is why PCI-e is way to go.  

Anyway, you need some decent diagnostics to see what is going on.  The manufacturer freebies won't dig deep enough.   You'll need to look at each disk, see if you are getting unrecoverable read errors, ECC errors, monitor link speed, and such as you pump data.  

I am biased to the one that is referenced with these example URLs relating to the SCSI transport info:  
There are products that will set you back around $1000 that will give you more info, but for $100 this one is a bargain.  Note it is not written for a consumer who doesn't know what they are doing, but not a lot of people even try to resolve link speed issues.  


Featured Post

2018 Annual Membership Survey

Here at Experts Exchange, we strive to give members the best experience. Help us improve the site by taking this survey today! (Bonus: Be entered to win a great tech prize for participating!)

Question has a verified solution.

If you are experiencing a similar issue, please ask a related question

Each year, investment in cloud platforms grows more than 20% (https://www.immun.io/hubfs/Immunio_2016/Content/Marketing/Cloud-Security-Report-2016.pdf?submissionGuid=a8d80a00-6fee-4b85-81db-a4e28f681762) as an increasing number of companies begin to…
New style of hardware planning for Microsoft Exchange server.
This tutorial will walk an individual through the process of installing the necessary services and then configuring a Windows Server 2012 system as an iSCSI target. To install the necessary roles, go to Server Manager, and select Add Roles and Featu…
In this video, Percona Solutions Engineer Barrett Chambers discusses some of the basic syntax differences between MySQL and MongoDB. To learn more check out our webinar on MongoDB administration for MySQL DBA: https://www.percona.com/resources/we…
Suggested Courses
Course of the Month3 days, 3 hours left to enroll

600 members asked questions and received personalized solutions in the past 7 days.

Join the community of 500,000 technology professionals and ask your questions.

Join & Ask a Question