SOLVED WARNING: You need to upgrade the volume TestPool

[scsi.king]

Hi, I am new to Freenas and just thought my older Seagate 500GB sata drive was going bad due to the same warning I just found in this conversation - so YES newbie's will think there is something wrong with my drive(s) based on this warning. I thus swapped the drive out with a Hitachi much newer 640GB hard drive and started continuing my test. My test is to dump 340GB of INSTALL's I have collected over the last 10 years, to this drive from my local 4TB hard drive using FreeFileSync for Windows. So far I have not got that warning to upgrade the pool; however, the Freenas 9.3 beta has fallen off line - or stalled out 3 times now. Twice on the older Seagate drive before and after this mornings update, and on the newer drive I switched it for.
What did I create:
Everything here was created with 9.3 beta - so no upgrade, no previous pools etc...
The first time I flagged the storage device (at creation time as a windows volume) and created the CIFS share after that.
The second time (after updating to an update this morning) I created the new pool on the newer drive, and told it was a UNIX based pool and then created CIFS share after that. I too became suspicious that the Windows attributes might be causing an issue (not bad for a newbie?) - especially because of my perceived issue with writing deep nested file structures (8 deep at times) and seeing the performance of ZFS drop to like less than 1 meg per sec when encountering these types of nestings.
In both scenarios, the single drive, single pool, just fell off line - to my windows system. The Freenas web manager stayed on line and acted like nothing had happened.
Hardware:
So I have a Windows 7 machine with 16GB Ram and set up my new Freenas box with a brand new ASUS MB H971 that has a built in NIC from Intel that only the Beta 9.3 can use, so I am using that so I can use the 16GB ram I put on this new board, where my previous 9.2.x Freenas only could take 8GB ram.
The 9.2.x Freenas software never fell off line, but I perceived it a little slow when traversing the file structures in order to sync a few changed file in the 400GB data sets I use... It was an AMD E350 low power motherboard with 8GB ram.
I have posted the graphs of the FreeNas system just after the last falling off line. It took around 15 min to occur, and approximately 50GB had been transferred out of 309GB to transfer.
Well after posting this, it fell off line again...
I will use an PCI-E Intel Nic, go back to 9.2.x and try all this again.
I hope this helps. Please contact me for any questions.

 

[Ericloewe]

Hi, I am new to Freenas and just thought my older Seagate 500GB sata drive was going bad due to the same warning I just found in this conversation - so YES newbie's will think there is something wrong with my drive(s) based on this warning. I thus swapped the drive out with a Hitachi much newer 640GB hard drive and started continuing my test. My test is to dump 340GB of INSTALL's I have collected over the last 10 years, to this drive from my local 4TB hard drive using FreeFileSync for Windows. So far I have not got that warning to upgrade the pool; however, the Freenas 9.3 beta has fallen off line - or stalled out 3 times now. Twice on the older Seagate drive before and after this mornings update, and on the newer drive I switched it for.
What did I create:
Everything here was created with 9.3 beta - so no upgrade, no previous pools etc...
The first time I flagged the storage device (at creation time as a windows volume) and created the CIFS share after that.
The second time (after updating to an update this morning) I created the new pool on the newer drive, and told it was a UNIX based pool and then created CIFS share after that. I too became suspicious that the Windows attributes might be causing an issue (not bad for a newbie?) - especially because of my perceived issue with writing deep nested file structures (8 deep at times) and seeing the performance of ZFS drop to like less than 1 meg per sec when encountering these types of nestings.
In both scenarios, the single drive, single pool, just fell off line - to my windows system. The Freenas web manager stayed on line and acted like nothing had happened.
Hardware:
So I have a Windows 7 machine with 16GB Ram and set up my new Freenas box with a brand new ASUS MB H971 that has a built in NIC from Intel that only the Beta 9.3 can use, so I am using that so I can use the 16GB ram I put on this new board, where my previous 9.2.x Freenas only could take 8GB ram.
The 9.2.x Freenas software never fell off line, but I perceived it a little slow when traversing the file structures in order to sync a few changed file in the 400GB data sets I use... It was an AMD E350 low power motherboard with 8GB ram.
I have posted the graphs of the FreeNas system just after the last falling off line. It took around 15 min to occur, and approximately 50GB had been transferred out of 309GB to transfer.
I wonder if the 2GB swap should be increased? Can it be increased at NAS install time? Is the SWAP on my 8GB flash drive I used to install the system? The graph shows it is maxed out in the reports.
I hope this helps. Please contact me for any questions.

I'm sorry, what?

That has absolutely nothing to do with this thread.

I recommend you read the plentiful documentation before try to hijack completely unrelated threads to your case.

 

[scsi.king]

I'm a newbie and got the same message you had. Someone said it was related to CIFS sharing which I use exclusively. How is that not related?

 

[Ericloewe]

You're running consumer hardware that is very much not recommended without ECC. You're running a single disk which is probably an old laptop model.

You yourself said you did not get the error with a new pool - which is as surprising as the sun setting later today. This means your situation has nothing to do with new feature flags.

 

[RoboKaren]

Agreed... And if you're a newbie to FreeNAS, you should be running 9.2.x -- not the beta. It's impossible for you (and for us) to disambiguate your newbie problems from any bugs that might be in the system.

 

[scsi.king]

jkh Administrator
SAID:
The reason for the warning is that we have actually observed, in the field, serious issues with the following very specific combination:

1. User has a pre-9.2 pool (usually 9.0.x) and 9.3 kernel.
2. User is doing a lot of CIFS sharing and uses extended attributes
3. User is manipulating a lot of files (we're not sure if directory creation/deletion is involved, or what) while extended attributes are also in use

This, after a certain period of time, results in pool corruption. We have tried like crazy to replicate this and have been entirely unsuccessful, which is why it is a warning and we're not posting long messages to the forum going "ZOMG!

1. I was only contributing to this thread because of the quote: "We have tried like crazy to replicate this and have been entirely unsuccessful". What does that mean to you in the context of trying to replicate the problem? I did not ask anyone to SOLVE my problems. I was offering data on something that may or may not help replicate the problem JKH was talking about.
2. I got the same message that started this thread about needing to upgrade my volume. I only mentioned that I was a newbie because someone mentioned newbie's might think they have to replace the drive due to this message - which I did.
3. I was coping a mass of install files using a CIFS share when the message appeared. So it's not just jkh's 1-3 assertion that this particular circumstance leads to this message. I did not have #1, and I was only copying files in #3
   
4. There is nothing wrong with either drive - they have now fully written the install folder with out issue as of going back to 9.2.x yesterday.

Good day.

 

[joeschmuck]

Since this is solved now and I recreated my pol just this morning using the current Beta and I didn't get this warning message again, I'll close this thread.