QNAP NAS Community Forum

by **astropolak** » Thu Jul 21, 2011 5:49 am

Is it common for posters to be allowed to wonder off topic on this forum? I thought the subject was " Smbd and Flush-9:0"

by **joelang1699** » Thu Jul 21, 2011 5:54 am

I was following advice on finding out Flush-9:0 kept showing up by running SMART check. This resulted in a dropped disk.

by **joelang1699** » Thu Jul 21, 2011 8:26 pm

Seatools says Disk 1 is now dead, RMA time!

Im still getting Flush in degraded mode!

by **florian.baumert** » Fri Jul 22, 2011 9:21 pm

As I had the same Problem on my 410 (flush-8:0 in my case and smbd) I had the theory what it has something to do with the filling degree (92%) of my disks. My suspect was that it write to the disks and that the checksum information is distributed so that reading and updating it blocks a long time. Alternatively there is to few cache to handle this without a lot of disk-io...

I "fixed" this by adding another 2TB drive which brings the "disk full" to 62%. Immediately the problem is not reproducable., I can write locally and over samba with full speed.

Seems like something QNAP needs to fix really urgent!

by **joelang1699** » Fri Jul 22, 2011 9:26 pm

You may be on to something there mate. My RAID 5 is at 85% in use and suffering from this issue. I deleted some stuff last week and the problem disappeared for awhile but when I start adding films again it was back.

by **joelang1699** » Fri Jul 22, 2011 11:43 pm

Okay i deleted some 300gb of info on my NAS. Flush:09 still causes spikes BUT i can now write to my NAS at full speed again. Storage size must be related to this issue then.

by **Anathema** » Sat Jul 23, 2011 12:43 am

I have similar problems on my TS-859 on which I am running RAID 6 with 8 disks (all are recommended enterprise 2 TB disks). Currently, the raid is 88% utilized - so the problem could be related to high capacity utilization as others have commented.

I think I had the same problem previously and got rid of it temporarily by resetting to factory settings, but the problem reappeared after a few weeks. I don't think I reduced capacity utilzation at the same time, but I may have so I am not sure what exactly removed the problem (it could have been deleting files as others have commented worked for them).

When I am NOT reading/writing to the RAID, smbd is running on the NAS, hogging 99.9% of the CPU. When I start a write to the raid, smbd shuts down and flush-9:0 takes over. It uses in the range of 5-10% of CPU, but write speeds are incredibly slow -- about the rate I could carve the data into stone with a toothpick

.

My theory has been that it is a write from Windows 7 to the raid that has gotten hung that the NAS seeks to sort out when it is idle, as the problem for me seems to have cropped up after heavy writing activity to the raid, but it could be just related to capacity as others have commented.

I love the QNAP NAS, but not in the crippled state that it is in now. I hope this is a priority issue for QNAP to fix, and I would like affirmation that the problem is being worked on, and suggested work-arounds in the meantime. At the moment I am backing up the raid to do another reset and update firmware -- if the last firmware update addresses this problem, a confirmation of that would be appreciated.

by **joelang1699** » Sat Jul 23, 2011 12:48 am

This is a bad sign if even the top of the range devices like yours are suffering from the same issue. The problem still exists when using the web file manger writing to the NAS so its not just Win 7. I also have the same issue with various Linux PC, Win Xp, Win Vista.

by **Anathema** » Sat Jul 23, 2011 12:53 am

joelang1699 wrote:This is a bad sign if even the top of the range devices like yours are suffering from the same issue. The problem still exists when using the web file manger writing to the NAS so its not just Win 7. I also have the same issue with various Linux PC, Win Xp, Win Vista.

Ye, using web file manager does write at incredibly slow speeds, writing from Windows 7 fails.

My hypothesis was that the problem was created from Windows 7 access, but then persists indefinitely -- I was thinking a failed read/write that never resets and instead remains in a "hung" state -- but that is just a blind guess.

by **florian.baumert** » Sat Jul 23, 2011 10:05 pm

Let me add an additional note. I saw two effects when my disks were full. One was the flush-9:0 using some cpu time and the awful slow write speed (if copying for example locally some files).

Additionally if I tried to write via samba, the smbd daemon took 99% CPU time. I straced it and smbd was stuck in a splice loop. Splice reads from one handle and directly writes to another. It always "succeddes" and writes zero (0) bytes. This way smbd uses all CPU without any progress.

Maybe this helps someone, I expect this to be a problem which is caused by but not related to the flush-9:0 (write) problem.

by **ivox** » Mon Jul 25, 2011 2:40 pm

joelang1699 wrote:This is a bad sign if even the top of the range devices like yours are suffering from the same issue. The problem still exists when using the web file manger writing to the NAS so its not just Win 7. I also have the same issue with various Linux PC, Win Xp, Win Vista.

Just the SAME!
The problem exists in all my computers, Macbook Pro, Windows XP PC and Win 7 PC.
Smb Afp and Webfilemanager, all write deadly slow.

by **ivox** » Mon Jul 25, 2011 2:44 pm

ivox wrote:
joelang1699 wrote:This is a bad sign if even the top of the range devices like yours are suffering from the same issue. The problem still exists when using the web file manger writing to the NAS so its not just Win 7. I also have the same issue with various Linux PC, Win Xp, Win Vista.

Just the SAME!
The problem exists in all my computers, Macbook Pro, Windows XP PC and Win 7 PC.
Smb Afp and Webfilemanager, all write deadly slow.

and my disk space usage: 76.7%
Available space: 1281.99 GB

by **ivox** » Tue Jul 26, 2011 1:16 pm

the problem is disk SPACE! or the Network Recycle Bin.
I deleted some files and emptied the Network Recycle Bin, then the write speed is normal now!
Is this good news for all of you?

by **schumaku** » Tue Jul 26, 2011 2:56 pm

ivox wrote:the problem is disk SPACE! or the Network Recycle Bin. I deleted some files and emptied the Network Recycle Bin, then the write speed is normal now!

Less files, and less storage space used - It does not matter where "apce" (generally speaking) is freed-up.

by **AirtjE** » Thu Aug 04, 2011 2:53 am

Encountered the same problem last few weeks. Was looking for the cause but just today I found out.
Seems indeed flush-9:0 ate up all CPU. Using a TS410 here with 3x 2TB (Seagate ST2000DL3) RAID5.

Couple of weeks back I saw the diskspace dropped to around 380 GBytes free. Still working OK.
Later I got a firmware upgrade from 3.3.x to 3.4.3. Since then I encountered the problem. Not sure if this is a co-incidence. Possibly also have something to do with free disk space dropped below 300 GB.
Ever since when I download large files directly to my NAS, or when I use uTorrent, connections to the NAS would drop suddenly and almost every program locks in my Windows 7 computer.

After reading this thread, I ended up cleaning my NAS from 270 GB free now to 337 GB free. Seems to be running fine now.
I was hoping to delay a new 4th disk until after the summer holidays but maybe its needed more than I thought...

QNAP NAS Community Forum

TS-419P smbd and flush-9:0 acting up

Re: TS-419P smbd and flush-9:0 acting up

Re: TS-419P smbd and flush-9:0 acting up

Re: TS-419P smbd and flush-9:0 acting up

Re: TS-419P smbd and flush-9:0 acting up

Re: TS-419P smbd and flush-9:0 acting up

Re: TS-419P smbd and flush-9:0 acting up

Re: TS-419P smbd and flush-9:0 acting up

Re: TS-419P smbd and flush-9:0 acting up

Re: TS-419P smbd and flush-9:0 acting up

Re: TS-419P smbd and flush-9:0 acting up

Re: TS-419P smbd and flush-9:0 acting up

Re: TS-419P smbd and flush-9:0 acting up

Re: TS-419P smbd and flush-9:0 acting up

Re: TS-419P smbd and flush-9:0 acting up

Re: TS-419P smbd and flush-9:0 acting up

Who is online