[SOLVED] ata1: SRST failed (errno=-16), boot extremely slow, BIOS disappears

I didn’t ask for assistance with this problem, but I want to post it anyway, for the benefit of others who encounter the same problem. In fact, I’m wondering whether we should advise forum members, as a general rule, to post problem solutions.

A few days ago, I replaced both the harddrive and the monitor in my 11-year-old AMD desktop. I then installed OpenSuSE 11.2 with Gnome 2.28.2 on the new harddrive. The installation went spectacularly well, and the images on my new monitor are glorious.

There was only one problem: Booting was extremely slow. Simply for BIOS to detect the presence of my harddrive would take over a minute. Grub would take another half minute to load, then issue several “ata1: SRST failed” complaints. I applied google to the message and found other people having extremely slow boots, but no one finding solutions. Some people blamed the video card, some blamed memory overload, and some blamed Western Digital.

I opened the box, looked at the new harddrive, and recalled my installation. The drive was one I had used briefly several years ago as a slave. Now that it would be the only drive in the system, it would be the master – so I had set the jumper accordingly.

Now, to confirm that the jumper was set correctly, I went to the Western Digital site and found the support documents. That is when I discovered that there is a third possible setting – a “single drive” that is NEITHER slave NOR master. Aha!

The single setting was indicated by the absence of a jumper, so I removed the jumper and taped it to the chasis, to make sure that I wouldn’t lose it. Then I restarted the computer and pressed “Del” to get the BIOS options.

Much to my dismay, BIOS indicated “no harddrive present”. What have I done?! Have I destroyed the drive? In a panic, I shutdown the computer without bothering to exit the BIOS screen. Then I opened the box and found the problem: I had forgotten to attach the lead from the power supply.

I attached the lead and closed the box. Confident that all was now well, I reattached all of the external cables and booted the computer. The monitor went crazy. Instead of the usual BIOS messages, I got a moving test pattern. Pressing “Del” had no effect. Once again, I pulled the plug quickly.

Have I destroyed BIOS by shutting down abruptly while in the BIOS option screen?! Have I pressed “Del” too many times?! I got out my BIOS manual and turned to the part of the trouble-shooting section that dealt with flashing a new BIOS into the system.

Finally calming down a bit, I wondered whether the new monitor might be fried or poorly connected. I put the old monitor back and booted again. Same test pattern!

This time, I patiently allowed the pattern to run its course. I noticed that the computer seemed to be going through the motions of a boot sequence. So maybe BIOS was still there after all!

I opened the box again and returned the jumper to its original position: “master”. Then I booted again. The test pattern disappeared, BIOS returned, and the boot completed, though slowly. All of my data was still there, on the drive! I counted my blessings a second time.

Now I went back to the Western Digital manual. This time I found a fourth jumper option: cable select. Ah, that sounds vaguely familiar! I moved the jumper to the CS position and booted one more time.

This time the boot proceeded at lightning speed, and the “ata1: SRST failed” messages were gone.

Problem(s) SOLVED!

Ah, the sweet mysteries of computing…

Everything that happens before grub loads is a hardware/firmware (bios or otherwise) issue.

The CS (cable select) setting in IDE drives is used to let the bios decide in which slot the drive will be seen according to the cable it is connected to. I never used this, it was buggy in the first mobos that supported it and using master/slave/single jumper was always easy. Are you sure the fourth setting isn’t the “single” option?

This test pattern is something I’ve never seen. It first I thought it was the monitor (as the “no/unsupported signal” jumping box you get onscreen when the signal frequency is wrong) but as it happened on two monitors it has to be some “feature” of your bios/motherboard. It might be worth to check the m/b manual.

With IDE going the way out, issues like this won’t happen anymore. I presume your son’s computer doesn’t have any sata ports? If it does, you may consider a sata drive for your next replacement. It’s one drive per cable, no jumper settings - except on the early ones, to select max supported transfer speed (150 or 300MB/s IINM).

Anyway, thanks for your report. It’s always interesting to hear about cases like this.

I just installed OpenSuSE 11.4 for x86_64 and I have this problem.

More specifically, I have a brand new (not even formatted yet) Western Digital Caviar IDE 500 Go. Until I read this thread, the hard drive was in Slave mode. Now, I placed the jumper on “Cable select”.

In both cases, I have the ata2: SRST failt (errno=-16) message at boot (before any other message by the way). However, my motherboard doesn’t have any SATA ports (hence, nothing is connected in sata in my computer).

The time isn’t much of a problem (it’s a server, so it shouldn’t reboot too often, and fast boot of OpenSuSE 11.4 works good). What is more of a problem is that the hard drive isn’t recognized by the OS, even though the BIOS does recognize it.

Let me note that this new hard drive is to be the slave, the master being another IDE drive. Concerning the cable connection, it leave the motherboard, goes to the DVD drive and ends on that new hard drive. This is the same hardware configuration I had before changing the aforementioned hard drive and it worked perfectly with 11.0.

Does any one have an idea concerning why my hard drive isn’t recognized and why there’s that message?

Thanks in advance!

On 2011-04-04 22:36, n0x wrote:
> Does any one have an idea concerning why my hard drive isn’t recognized
> and why there’s that message?


I think you should report the issue in Bugzilla.

However, The other day I read this - I don’t know if it is the case:


> Problems with older hardware
> Hard disk and / or CD not found during installation
> Some older systems (typically Pentium-I 75-150 MHz) need the ‘ide-generic’ module loaded before the Hard disk and / or CD can be detected by the installer. If you encounter such behaviour, try re-starting the installation with the following added to the boot line:
> ‘insmod=ide-generic’

Cheers / Saludos,

Carlos E. R.
(from 11.2 x86_64 “Emerald” at Telcontar)

Thank you for this message. It reminded me that the 2 SATA (samsung) drives have jumpers. After some googling I found that only one of the available jumpers does something using this article: Tech ARP - Samsung EcoGreen F2 (HD154UI) 1.5 TB Hard Disk Drive Review.

In my case a drive was not always recognized and the drive detection was kinda slow. I found out that my motherboard had a 1.5G SATA driver and that a jumper could be placed to limit the drive to 1.5G in case the driver could not handle them.

Now it functions properly, so thank you!

This worked for me on CentOS 6.3 as well. The system wouldn’t boot after a kernel update and I was getting the same errors. I changed the jumper on my WD IDE drive and poof! Fast boot!


I wanted to say thanks as this post saved me a lot of time.
I put an old hard drive in my main computer and changed the jumper from “slave” to “master”. Boot time was terrible and I got the “SRST Failed” error. When I eventually was able to log in, I changed /etc/fstab to mount the drive automatically and rebooted. It was worse… I ended up in some secure mode, went back to the old /etc/fstab and rebooted. This time I could eventually log in, searched for “SRST failed”, found your post, shut down, changed the jumper, restarted… and all is well. It boots as fast as without the new old drive.