Hi everyone!  What the hell is this - "jnl: unknown-dev: flushing fs disk buffer returned 0x5"? Cel D 2,66Ghz, 1480Mb DDR2, 128Mb ATI Radeon x200, Realtek AC97, 80Gb Maxtor HDD on primary master.

I'm getting this exact same error:


disk0s2 is an HFS+Journaling partition on an internal PATA used for storage. Main startup disk is a SATA 120 GB (Full specs in sig).
I have deleted and formated the volume over and over. Now i'm thinkin about deleting the whole disk (it also contains a WinXP install) to see if that fixes it. Ironically, same 80GB Maxtor on Primary Master as you screamroad:



What are you full system specs screamroad?

I dont remember seeing this error all that long (routinely check system.log). A Google search indicates the problem tends to happen with XServers with RAID devices. Others with this problem also report a 15-20 sec delay accessing files on the volume, but I dont get this delay. The drive works as perfect and fsck_hfs always reports it good. Yes, disabling Journaling will prevent the error, but I have frequent power outages and would like to have Journaling enabled.

It appears to be a drive specific issue according to here. If you're having this problem too, please post the HD make and model!
Alsouseful..

Any clues at all what could be causing it?? More important, is it harmless?

Thanks for any guidance,

This is very interesting.

I have a MacPro. It had 3 x 500G hard disks set like this:

Alpha Quadrant (Bay 1) Système
Beta Quadrant (Bay 2) iTunes and iPhoto
Gamma Quadrant (Bay 3) Sources, Movies, VMware Virtual machines.

Plus an iBook Studio External 1T drive from Western Digital, plugged into the FireWire 800.

6G of RAM.

I just upgraded to 4 x 1.5T drives (Alpha, Beta, Gamma and Delta), added another 4G of RAM for a total of 10G.

Since this upgrade, I have iTune that stop to play for 10 to 15 secondes sometimes. I thought first it was a bad drive. Seeing that Alpha is my system drive, I did this drive at the end. before that swap, everything was fine. So I swap Alpha and Delta reformatting them. Still there. I even sequentially take out one drive, switch back one old 500G, remove memory and today, I tried to unplugged the FireWire 1T external drive. Still the problem is there. I just thought this morning to take a look into the system log and found that exact error message at the specific time it happen this morning.

Then I filter only this specific message and found a lot of lines that look similar to the times when I got iTune that stopped playing.

And it is always on disk1s2 : my iTune disk !

Nov 2 07:45:20 VoieLactee kernel[0]: jnl: disk1s2: flushing fs disk buffer returned 0x5
Nov 2 08:33:42 VoieLactee kernel[0]: jnl: disk1s2: flushing fs disk buffer returned 0x5
Nov 2 08:51:41 VoieLactee kernel[0]: jnl: disk1s2: flushing fs disk buffer returned 0x5
Nov 3 08:49:07 VoieLactee kernel[0]: jnl: disk1s2: flushing fs disk buffer returned 0x5
Nov 3 09:20:55 VoieLactee kernel[0]: jnl: disk1s2: flushing fs disk buffer returned 0x5
Nov 3 09:31:40 VoieLactee kernel[0]: jnl: disk1s2: flushing fs disk buffer returned 0x5
Nov 3 13:00:39 VoieLactee kernel[0]: jnl: disk1s2: flushing fs disk buffer returned 0x5
Nov 3 13:05:43 VoieLactee kernel[0]: jnl: disk1s2: flushing fs disk buffer returned 0x5
Nov 3 13:06:44 VoieLactee kernel[0]: jnl: disk1s2: flushing fs disk buffer returned 0x5
Nov 3 13:09:45 VoieLactee kernel[0]: jnl: disk1s2: flushing fs disk buffer returned 0x5
Nov 3 17:05:53 VoieLactee kernel[0]: jnl: disk1s2: flushing fs disk buffer returned 0x5
Nov 3 22:46:04 VoieLactee kernel[0]: jnl: disk1s2: flushing fs disk buffer returned 0x5
Nov 4 07:41:10 VoieLactee kernel[0]: jnl: disk1s2: flushing fs disk buffer returned 0x5


I am thinking to swap this disk back to it's 500 MB just to see if it could be this drive that is defective...

Any thought ?

Yves




Hey guest what?

Just about 5 minutes after having post this above, iTune stop to play again, and came back at exactly 8:26:28 , I did a refresh in the window log that was still opened:

Nov 3 17:05:53 VoieLactee kernel[0]: jnl: disk1s2: flushing fs disk buffer returned 0x5
Nov 3 22:46:04 VoieLactee kernel[0]: jnl: disk1s2: flushing fs disk buffer returned 0x5
Nov 4 07:41:10 VoieLactee kernel[0]: jnl: disk1s2: flushing fs disk buffer returned 0x5
Nov 4 08:26:28 VoieLactee kernel[0]: jnl: disk1s2: flushing fs disk buffer returned 0x5


Isn't it interresting. All my 4 drives are the Seagate 1.5T that the link from the above reply is talkin about!!

Hi Yves,

I'm glad my link was helpful. I spent quite some time and scoured Google to research anything I could about that error.
It appears to be a problem with certain hard drives themselves, such as my Maxtor 6Y080L0 80GB and your Seagate 1.5TB (which the person in my link also prob has).

The question is, is this a harmless incompatibility between HFS+Journaling and our specific drives or a symptom impending drive failure. My Maxtor 6Y080L0 is a few months old and I dont remember seeing this error the whole time in my system.log. Yet you tested it with 4 brand new Seagate 1.5TB units and all got the same thing, correct?? Unless you somehow picked up 4 bad drives (!!), it looks like an incompatibility in how HFS+Journaling tries to do its business. Yves, try disable Journaling one of your 1.5TB's and see if the error goes away for that disk (hold down option FIRST and go to File > Disable Jornaling in Disk Utility).

Update: But the problem also appears to be more than OS X HFS+Journaling specific. According to this post, its affecting Ubuntu and Windows users too. 8 pages their of this same weirdness when it comes to journaling and flushing disk cache operations. According to that long thread, Seagate is saying its an OS issue (bug in Linux kernel, Windows drivers), or a chipset issue (cant access drive so big as 1.5 TB). But i have a small ole 80GB Maxtor, which I wonder if they have any firmware update for...

Anyone else with this error? What model drive do you have??

THANK GOD!!

after 2 days of searching and re-installing, I tried "site:insanelymac.com 1.5T" and bam..

2x brand new Seagate 1.5T, same 5-10 seconds of random silence

did you solve this?

edit: sorry, it looks like everyone is waiting for the firmware update..

Well i'm glad i was able to solve (or at lesat direct to...) everyones problem...

But I still have the same problem for myself (flushing fs disk buffer returned 0x5), except i dont have a Seagate 1.5TB drive.....

Anyone else?

I have the same prob.
We have also the Mac Pro (MacPro1,1 - first Intel Mac Pro), I just upgraded it with the Seagate 1.5TB hard drives, now I have this internal setup:
- 1x 320 GB Seagate ST3250824AS (system)
- 2x 1.5 TB Seagate ST31500341AS (music and music backup)
- 1x 750 GB Seagate ST3750640AS (itunes and video)

When using Logic Pro 8 it hardly happens, but it happens a lot when playing even a single MP3 from the 1.5 TB disks.. like every five minutes? annoying.

Once i noticed trouble like this (but didn`t checked the system log, if it`s the same prob, but seems similar) on our newer Mac Pro (early 2008 model - single quad-core), equipped with 1 TB Seagate drives. And it was much more deeper. On this older Mac Pro I can do nearly everything else, when the computer freezes the playback, but on the other Mac Pro it also disabled my Wacom Bamboo tablet (other USB equipment was working flawlessly) and worse - it happened when trying to record into the Logic Pro 8..

Hope some solution will come up early..

Hi Aluminum Sponge,

I have the solution, and the explanation!

The solution is not total, you'll see why. But here is the explanation...

First, you can use the Disk Utility to transform an HFS+ only partition and make it become Journalized HFS+ , but not the inverse, cause the button is shadowed!

To resume my setup:

MacPro 2008 (2 x 2.8 Ghz Quad-Core Intel Xeon)
10 GB 800 Mhz DDR2 FB-DIMM of RAM
4 eSATA disks of 1.5 TB (Seagate ST31500341AS)

I cannot believe I have to wait as long as 15 seconds after a system such as this in 2008! That is absolutely innaceptable.

IMHO, this problem is NOT a drive problem. Although I have to exchange one bad drive on warranty recently, the issue wasn't a faulty drive because I fixed the problem before the drive failed!

My issue started when I changed my 3 Seagate disks of 500 GB each and replaced them with 4 Seagate disks of 1.5 TB. To be honest, I didn't came back here to read the thread after I had fix my problem. (iTunes stop playing for about 10 to 15 seconds). But since yesterday, iDVD started to slow down and I am back.

Again, IMHO, the problem is the OS, more precisely, the Journaling. When you format a drive, never choose HFS+ Journalized, use HFS+ only instead!

Why?!

With HFS+ Journalized, like you already mentioned, the OS keep track of the progression of files I/O in a buffer. It does that in case of a crash, like when you lost electricity. In such a case, it would be able with its journal to recover the file I/O where they where before the crash. But the downside is that at some point, when its buffer is full, it need to flush it, meaning pushing the buffer content onto the real physical disk sectors. And this consume I/O and sucks!

I just don't understand why I didn't have this problem before my upgrade. Is it because I have 4 disks instead of 3? I have doubts because in a MacPro, each bay have its own controller. That is one reason of its good performance. By the way, if you have more than one drive, you will get better performance if you spread your Input/Output on different drives. Like in Final Cut, put the movies you read on one drive and save the movies you create on another drive, that should boost your Input/Output performance in theory.

Or is it the Seagate specific 1.5TB model? That could make sense seeing that the line in the log file report "unknown-dev ". Maybe we should read "unknown firmware" ! I will be able to tell that in a couple of days. See details below.

In order to fix my iTunes stop playing music problem, I simply decided to reformat all my drives without Journaling. My iTunes Library is on my second drive "Beta Quadrant". Since this decision, I NEVER got any cut anymore in my music. I even have checked the log file for flushing activities and found nothing. I was sure I did fix the problem, but it was partial.

Unfortunately, Mac OS X doesn't let you format the system drive without journaling!

So I finished out my setup with this scenario:

Bay 1: Alpha Quadrant HPFS+ Journalized

OS X and Applications

Bay 2: Beta Quadrant HPFS+

Libraries and Projects only files (iTunes, iPhoto, Aperture, iDVD, Final Cut Express)

Bay 3: Gamma Quadrant HPFS+

All sources files (MOV, miniDV, DVD, etc)

Bay 4: Delta Quadrant HPFS+

All movies files done with Final Cut Express

Since this setup, I never found any other flush messages in my log files and iTunes run without any problem. Still I got some delay sometimes when I was using Final Cut Express, but I always thought it was because the drives was starting to spin again after having fall down to sleep because of no activity on them.

Now... Yesterday I started to use iDVD to create my Christmas gifts and I realized that the multicolor mouse cursor was spinning very frequently. I decided to monitor the log again and found out that I had the exact same problem of Input/Output files flushing but this time it was only affecting iDVD. Of course, iDVD use a lot of I/O in order to let you visualize the result dynamically.

Realizing that all applications was installed on the system disk, the one that was still formated in Journalized HFS+, I tried to copy the iDVD application package (folder) onto another non journalized HFS+ formatted drive and run it from there. Same problem, but sill reported on the Bay 1, the system drive formatted with Journalized HFS+.

My understanding of this is that these I/O consuming applications must use some library files under the system drive that make them use a lot of I/O even if they are run from another drive, as an example, all themes of iDVD are saved under "/Library/Application Support/iDVD/Themes"

Seeing that Mac OS X doesn't let us format the system disk without the Journalized HFS+ setting, and seeing that I never had this issue before my upgrade, I will make a last test. I will use one of my old 500 GB disk drives and use it for my system disk. If I don't have this problem anymore, I will conclude that it is drive model specific. In the contrary, I will have to admit that Mac OS X have more difficulty to handle 4 controllers instead of three!

I will keep you update...

P.S.: I bet that a lot of peoples have the same issue but never realized it. Those of you who are using Mac OS X, I strongly suggest you to try this to confirm it:

1- Open a Terminal window

Applications -> Utilities -> Terminal.app

2- At the prompt, type this line ( you can cut&paste ) :

clear; tail -n 5000 -f /var/log/system.log | grep "flushing"

The "clear" command will clear out the terminal screen and put your cursor at the top. the "tail - 5000" command will list the last 5000 lines of the system log files ( -f /var/log/system.log ). And the " | grep " (read "pipe grep") command will take care to filter out everything except the lines that include the string "flushing". After hitting the return key, the cursor will stay at the end and it will output any new line with the string "flushing" until you cancel it with the "CTRL-C" keys sequence.


3- Then, open iDVD and play with it. As soon as you obtain a delay with the multicolor mouse pointer spinning, wait until the mouse cursor came back as normal (I bet 10 to 15 seconds) and take a look at the terminal log screen. I bet you will read something similar to :

Dec 20 08:52:08 VoieLactee kernel[0]: jnl: unknown-dev: flushing fs disk buffer returned 0x5

Please, let us know your results

See you all,

Yves

Hello

I have a good news, I found the way to reformat the system drive without journaling.

As explained in my last reply, I wanted to use my old 500 GB drive to see if the problem would still be there. So I replaced my "Delta Quadrant" drive with the 500 GB drive and I partitioned it with one partition named "Alpha Quadrant 500 GB". I specifically choose extended HFS+ without journaling. I was expecting anyway that the transfer of my original "Alpha Quadrant" over it would have make it a journalized one, as it happen the last time.

But it didn't! I don't remember if the last time I used the partition tools first but it seem that if you first choose to partition the drive, it would let you choose the non journalized extended HFS+ file system.

Then I use the "Restore" tab of the "Disk Utility" tool. First you pick any drive from the list. Then you click on the tab "Restore". With your mouse, you drag the source drive into the "Source :" field and you drag the destination drive into the "Destination :" field.

When you click the "Restore" button, don't erase the content of your destination drive, this way it will keep your file system choice when you partionned it.

I choose my 1.5 TB "Alpha Quadrant" journalized HFS+ drive as my source and pick the 500 GB "Alpha Quadrant 500 GB" non journalized HFS+ drive as my destination.

I finished with an exact 1 for 1 copy of my system drive excepted that this new one wasn't journalized.

Then I choose this new drive as the boot drive in the Preference and reboot.

Bam! Problem fixed!!

After a couple of tests, I decide to do the exact same procedure this time copying the 500 GB drive over the 1.5 TB and reboot again with the latest as the boot drive. I ran Final Cut Express exporting a movie, iMovie HD exporting a movie and played with iDVD all at the same time.

NO MORE FLUSHING MESSAGE in the log file and the system was running fast and smoothy!

Today, I want to push further my test. I will put back the 500 GB "Alpha Quadrant 500 GB" drive as my system drive, but I will convert it in a journalized file system, an test...

See you later.

Yves

So I did it.

I put back my old 500 GB "Alpha Quadrant 500 GB" drive into the MacPro, make it a journalized file system and ran Final Cut Express, iDVD, iTunes, Mail, Starry Night Pro Plus, all at the same time.

I played with iDVD as usual and never got the "flushing" error message into the log file!!!!

So my conclusions are that it is caused by the journal daemon when you choose the HFS+ Journalized file system, but only with specific drive.

The Seagate 1.5 TB model ST31500341AS is one that have the problem. I have 4 disks of this type in my MacPro and they all have the problem when formatted with the journalling enabled.

The Seagate 500 GB model ST3500630AS doesn't have the problem. I had 3 disks of that type in my MacPro before my upgrade and I never sense any slow speed at that time and they were all formatted with journaling enabled.

I even make the test with my MacBook which I recently upgraded its 80 GB disk with an Hitachi 320 GB disk and I didn't got the "flushing" error message, never! Yes it is formatted with journalling enabled.

The question now is "Who is responsible ?"

The Journalized file system daemon that doesn't recognize the Seagate 1.5 TB or the Seagate 1.5 TB that is not compatible with the file system?!?

Anyway, the solution is to use the Extended HFS+ NON JOURNALIZED file system when you format this drive. And make sure that you use the partition tool to partition it before you format it.

Best regards all,

Yves

Firmware update for 1.5T Seagate is out

check http://enl.iten.me/technology/update-firmw...m-halt-bug.html

This looks good for a PC, but I own a MacPro. I guest that would imply taking out all my drive but one, one after the other...

I tried to send an email to the address this article was giving, we'll see Seagate answer...

Thanks anyway

Yves