Error - Assuming Drive Cache: Write Through

AceRoom · #1 8th December 2011, 05:17 PM

Code:

[  420.590676] sd 5:0:0:0: [sdb] Assuming drive cache: write through
[  472.818201] sd 5:0:0:0: [sdb] Test WP failed, assume Write Enabled
[  472.819804] sd 5:0:0:0: [sdb] Asking for cache data failed
[  472.819813] sd 5:0:0:0: [sdb] Assuming drive cache: write through

I keep getting this error on dmesg. It also shows up in the virtual terminals. I believe that it is probably caused by the internal card reader as I don't have any other devices connected (Unless the internal cdrom is counted).

Is this an issue? If so, can anyone help me to rectify it. TIA

agillis · #2 10th December 2011, 08:44 PM

I'm getting this as well. sdb is my DVD drive. Happened after an upgrade from Fedora 14 to 16. Could this be a bug in F16?

stevea · #3 11th December 2011, 12:54 AM

It's not an error. It's describing a detection test of the drive cache -
It tries write-protect then assumes write-enabled, asks for cache data (failed) so assumes cache it write-thru.

Errors are generally things clearly identified w/ the word "error" . [I'm not being snarky - it is hard to tell]

agillis · #4 12th December 2011, 02:43 PM

OK thanks, Do you know if there is any way to disable this to keep my console screen from filling up?

sidebrnz · #5 12th December 2011, 09:14 PM

I've been seeing that too on my desktop, when I insert a USB key. However, if I have two CLI sessions open, and one of them's root, it only appears on the non-root session, even if I'm working on the root one. Odd.

AceRoom · #6 14th December 2011, 04:35 AM

Yeah, any way to disable it from showing this would be very useful. It keeps interrupting when using a virtual terminal session...

Netlurker · #7 16th December 2011, 12:08 PM

I am seeing this in Fedora 16 as well. With the most recent kernels it seems to stop the boot process however. If I fall back to 3.1.4-1 I can boot, but still see the errors in syslog. With kernels 3.1.5-1 or 3.1.5-2 it never seems to be able to boot just hanging at the message which repeats every couple of minutes. I have left it 20+ minutes to ensure it doesn't eventually come up but just see the message over and over. In my case it's sdb which seems to be a multi-card reader on the USB bus (Bus 001 Device 002: ID 0bda:0158 Realtek Semiconductor Corp. USB 2.0 multicard reader) on an Acer Aspire 4530.

Update: Turns out the whole thing was a red herring. The messages are just what would normally show up in /var/log/messages and are apparently harmless. The appearance of a hung boot was actually a problem with the proprietary nvidia driver. I was using the kmod version which doesn't deal with kernel updates well. I am now using the akmod version which is supposed to be more graceful about that. To verify if this is your issue, edit the grub config for the kernel you are trying to boot with adding a 3 to the end of the kernel line so it doesn't try to start X. Then login and run startx, if there is a problem you should see it on the console. In my case I ended up just doing a "yum remove kmod-nvidia*" and then "yum install akmod-nvidia" as I already had the RPMFusion repo's enabled.

sidebrnz · #8 27th December 2011, 12:14 AM

Currently, I'm getting those messages during boot on my laptop, running F16, if and only if I boot into an F16 kernel and booting into runlevel 3 doesn't help. It just keeps spewing out those messages, and never completes boot. If, however, I boot into the old F14 kernel, it boots just fine. And, nVidia drivers aren't an issue here because I don't have nVidia graphics on my laptop. I'm not sure what's on sdb, but I don't boot with a CD or flash drive loaded and although it has a card reader, I've never used it. Any more interesting ideas?

marriedto51 · #9 28th December 2011, 09:24 AM

Well, yes, I see this message too. It is (as has been said) harmless (apart from interrupting the flow of your typing) and originates with the kernel. At a guess it is a mild bug in that the message has not had the right level of "urgency" assigned to it in the Fedora builds of the kernel.

sidebrnz · #10 28th December 2011, 05:44 PM

I don't call something that (in my case) stops the boot process dead in its tracks "mild." If you want to add to the Bugzilla report at https://bugzilla.redhat.com/show_bug.cgi?id=769747 it might even help.

marriedto51 · #11 28th December 2011, 06:27 PM

Quote:

Originally Posted by sidebrnz

I don't call something that (in my case) stops the boot process dead in its tracks "mild."

Fair point. I should have read the earlier posts more closely. (I just get the messages, but no problem with booting.)

linker3000 · #12 21st February 2012, 03:43 PM

I got this error - which stopped my laptop from booting - following a combination of a kernel update and then an update to VirtualBox to support the new kernel.

I suspect I messed up the VirtualBox update and ended up with some parts from the OSE version (manually downloaded in the dim and distant past) and some from the Oracle yum repository.

I fixed my install by booting into non-GUI mode by adding 3 to the end of the kernel line during GRUB booting and then doing:

yum remove VirtualBox*
yum install VirtualBox-4.1

After that the messages disappeared and the laptop booted normally.

sidebrnz · #13 21st February 2012, 06:28 PM

AFAIK I don't have VirtualBox installed. Next time I've got time to get my laptop updated, I'll check.

dharper · #14 2nd March 2013, 06:57 PM

So I keep getting this message as well, and only since I updated to Fedora 18. I was running 13 on my laptop before and never saw it. Anyone figure out how to stop this message from spamming the CLI? I even get the message output when I am working in VIM ...

sidebrnz · #15 2nd March 2013, 07:12 PM

I eventually booted from an F 17 LiveCD and found that it worked fine. I decided that this meant that there was something wrong with my system that the newer kernels didn't like so I bit the bullet and did a clean installation of F 17, which got rid of the problem. Considering that you (apparently) went from 13 to 18 in one giant leap, you may want to consider doing the same thing.