How to check for Write Errors?

[JeffWScott]

I have a Ubuntu server which runs some software to download files to the QNAP's storage and then "unzips" those files into 1 big file (about 10-15GB). Then there is a completely different piece of software, running on the same Ubuntu server, that moves that file from one QNAP shared folder to another after is has been unzipped.

The Ubuntu Server is using NFS to mount the Qnaps Shares.

The first processes occasionally is reporting a write error while unzipping. If the first process completes fine the second process occasionally reports a write error while moving the file from location a to location b.

My HDD smart is reporting that everything is Good. So how do I go about diagnosing what the problem is?

NAS: QNAP 219-p
Firmware: 4.1.0

This issue cropped up about 2 months ago.

Thanks!

[pwilson]

I have a Ubuntu server which runs some software to download files to the QNAP's storage and then "unzips" those files into 1 big file (about 10-15GB). Then there is a completely different piece of software, running on the same Ubuntu server, that moves that file from one QNAP shared folder to another after is has been unzipped.

The Ubuntu Server is using NFS to mount the Qnaps Shares.

The first processes occasionally is reporting a write error while unzipping. If the first process completes fine the second process occasionally reports a write error while moving the file from location a to location b.

My HDD smart is reporting that everything is Good. So how do I go about diagnosing what the problem is?

NAS: QNAP 219-p
Firmware: 4.1.0

This issue cropped up about 2 months ago.

Thanks!

Have you tried using the "dmesg" command on both your Ubuntu box, and/or the NAS?
How about checking the logs on both?

Sorry, I can't help with your NFS issue myself. I abandoned NFS over 20 years ago, in favour of Samba on all my Linux devices/systems.
I've been running Linux for 22 years. I have 3 Linux computers, plus more than half a dozen embedded Linux devices, and I don't use NFS at all.

Have you attempted to test your undisclosed Drive Make/Model with the undisclosed manufacturers diagnostic tools?

Why are you "moving" files (twice no less) via NFS anyway? If the file starts on the QNAP, and ends on the QNAP, why aren't you copying the file locally on the NAS itself. You can execute commands on the NAS via SSH directly from your Ubuntu box command line this way, without clogging up your network bandwidth by needlessly moving the file via NFS (twice).

For example:

Code: Select all

ssh admin@qnap mv /share/Public/Folder1/File1 /share/Public/Folder2/File1

(assuming your NAS is known as "qnap" on your network)

Setup SSH Authorized Keys, if you don't want it prompting for password when you execute this command on your Ubuntu box, for example if executing this command from a cron job, or init script, or even personal script.

Your method is a complete waste of your network bandwidth, and moving a large 10-15GB file via NFS across your network twice when you don't need to move it at all, strikes me as a very poor implementation.

[JeffWScott]

Have you tried using the "dmesg" command on both your Ubuntu box, and/or the NAS?
How about checking the logs on both?

Sorry, I can't help with your NFS issue myself. I abandoned NFS over 20 years ago, in favour of Samba on all my Linux devices/systems.
I've been running Linux for 22 years. I have 3 Linux computers, plus more than half a dozen embedded Linux devices, and I don't use NFS at all.

Have you attempted to test your undisclosed Drive Make/Model with the undisclosed manufacturers diagnostic tools?

Why are you "moving" files (twice no less) via NFS anyway? If the file starts on the QNAP, and ends on the QNAP, why aren't you copying the file locally on the NAS itself. You can execute commands on the NAS via SSH directly from your Ubuntu box command line this way, without clogging up your network bandwidth by needlessly moving the file via NFS (twice).

For example:

Code: Select all

ssh admin@qnap mv /share/Public/Folder1/File1 /share/Public/Folder2/File1

(assuming your NAS is known as "qnap" on your network)

Setup SSH Authorized Keys, if you don't want it prompting for password when you execute this command on your Ubuntu box, for example if executing this command from a cron job, or init script, or even personal script.

Your method is a complete waste of your network bandwidth, and moving a large 10-15GB file via NFS across your network twice when you don't need to move it at all, strikes me as a very poor implementation.

Whoa! Great response, thank you!

I ran dmesg on the NAS and did get some output. If "nfsd" is related to my NFS shares then this would be a good indication that NFS is causing the problem. Do you agree?

Code: Select all

[1679551.039266] nfsd: non-standard errno: -14
[1746370.160512] nfsd: non-standard errno: -14
[1773337.475526] nfsd: non-standard errno: -14
[1779324.822199] bad rx status 2f84c801, (crc error)
[1805892.085498] nfsd: non-standard errno: -14

(FYI my HDDs are Western Digital RED 2TB)

Can you please point me towards the logs you are referring to?

The software cannot be altered so I cannot change the way it unpacks things. I can only give it locations and let it do its thing. I have it setup the way it currently is because the QNAP is not powerful enough to run both pieces of software, and the Ubuntu Server does not have the HDD capacity to hold the files.

Kind of a catch 22.

I remapped all my mounts to Qnap using Samba (cifs). I'm going to run like that for awhile and see how that goes.

Your expertise on the matter is refreshing.

Thanks again!

[pwilson]

Have you tried using the "dmesg" command on both your Ubuntu box, and/or the NAS?
How about checking the logs on both?

Sorry, I can't help with your NFS issue myself. I abandoned NFS over 20 years ago, in favour of Samba on all my Linux devices/systems.
I've been running Linux for 22 years. I have 3 Linux computers, plus more than half a dozen embedded Linux devices, and I don't use NFS at all.

Have you attempted to test your undisclosed Drive Make/Model with the undisclosed manufacturers diagnostic tools?

Why are you "moving" files (twice no less) via NFS anyway? If the file starts on the QNAP, and ends on the QNAP, why aren't you copying the file locally on the NAS itself. You can execute commands on the NAS via SSH directly from your Ubuntu box command line this way, without clogging up your network bandwidth by needlessly moving the file via NFS (twice).

For example:

Code: Select all

ssh admin@qnap mv /share/Public/Folder1/File1 /share/Public/Folder2/File1

(assuming your NAS is known as "qnap" on your network)

Setup SSH Authorized Keys, if you don't want it prompting for password when you execute this command on your Ubuntu box, for example if executing this command from a cron job, or init script, or even personal script.

Your method is a complete waste of your network bandwidth, and moving a large 10-15GB file via NFS across your network twice when you don't need to move it at all, strikes me as a very poor implementation.

Whoa! Great response, thank you!

I ran dmesg on the NAS and did get some output. If "nfsd" is related to my NFS shares then this would be a good indication that NFS is causing the problem. Do you agree?

Code: Select all

[1679551.039266] nfsd: non-standard errno: -14
[1746370.160512] nfsd: non-standard errno: -14
[1773337.475526] nfsd: non-standard errno: -14
[1779324.822199] bad rx status 2f84c801, (crc error)
[1805892.085498] nfsd: non-standard errno: -14

(FYI my HDDs are Western Digital RED 2TB)

Can you please point me towards the logs you are referring to?

The software cannot be altered so I cannot change the way it unpacks things. I can only give it locations and let it do its thing. I have it setup the way it currently is because the QNAP is not powerful enough to run both pieces of software, and the Ubuntu Server does not have the HDD capacity to hold the files.

Kind of a catch 22.

I remapped all my mounts to Qnap using Samba (cifs). I'm going to run like that for awhile and see how that goes.

Your expertise on the matter is refreshing.

Thanks again!

Check in /var/log on both devices.

[JeffWScott]

I would like to report that I have not had any issues since switching over to Samba mappings.

Thanks for your help pwilson!!