I have a 4 disk setup running on Unraid 5.0.6 Pro.  Parity with 3 data disks.  I recently deleted some files from my data disk1 and am trying to copy new large files to this drive.  The file starts to copy but connectivity drops in the middle of the transfer.  I've swapped out the switch, cable,and have even tried from a different computer/os.  I attached a screenshot showing the ping of what happens while the file is transferring mid fail.  I was pinging the host name not the direct IP address.  I am configured for DHCP not static.  I find it very strange that it changes from pinging 192.168.1.55 to 192.168.1.254 in the middle of the transfer.  It almost seems like Unraid is having the issue, not my network. 

 

I only seem to have this problem when putting data on drives that have previously had data on them.  If i try to copy to my newest drive that has not had files deleted from it this problem does not present itself.

 

Can someone with more experience shed some light on what may be happening?  My thoughts are it has something to do with parity having to recalculate on the drive that once had data on it and it may be causing a memory leak or buffer overflow in Unraid.

 

Any help will be appreciated.

Thanks in advance.

What is the DHCP lease time in your router?

My DHCP lease time is 60 days.  I did restart the unraid box as one of my first troubleshooting steps, but forgot to mention that in my first post.  As far as I can tell there are not IP conflicts on the network.  I also failed to mention that i did a Parity check also wich found no errors.

I'm having the same problem. I just set up my server, both disks are new. In order to get my transfer speeds up, I added a gigabit switch and had to change the jumbo frame sizes (on both my home PC and my unRAID box) to 9000 mtu. This increased my throughput to ~100MB/s, but then I have the same connectivity issues as OP.

 

If I transfer at 11MB/s, I can transfer without losing connection (14 hours now and running).

 

Is it possible that the data is moving TOO fast? Should I throttle the jumbo frame sizes a bit?

 

NOTE: For the initial load, I've deactivated the parity drive - this is just a straight copy across the network.

OK so after some thought I decided to ping the ip address as well as the host name.  The ip address and host name both stopped responding at the same time but the ip address ping did not change over to the different address.  From there I found out that the computer I was copying from was the 192.168.1.254 ip address and windows pings it's own ip address\gateway after getting no response from the network for a period of time.  So that part of my post is solved.

 

So now i know that the unraid box is falling off the network in the middle of the transfer. I do not have this problem while playing files back or copying files to other drives.  It only seems to happen on the drive that i freed up space on and then tried to copy files back onto that drive.

If it helps...

 

This error occurs at random times - sometimes right at the beginning of a copy, sometimes four files in. When it occurs, I can't delete the partially created file as it registers as in use. I must stop the array, restart it, and then I can delete the file.

 

I tried cutting the jumbo packet size back to 6000, but it's still occurring.

If it helps...

 

This error occurs at random times - sometimes right at the beginning of a copy, sometimes four files in. When it occurs, I can't delete the partially created file as it registers as in use. I must stop the array, restart it, and then I can delete the file.

 

I tried cutting the jumbo packet size back to 6000, but it's still occurring.

 

Sounds like you may be having a different issue than I'm having.  I do not have to stop the array and the file deletes itself as soon as I stop the transfer as long as the share has reconnected.  I'm not getting any file in use messages.  I do not have to restart the array to try sending the file again.  Although it does not always happen at the same time in the file.  Sometimes it happens around 25% of the way in sometimes it happens when the file is almost finished.

OK so after some thought I decided to ping the ip address as well as the host name.  The ip address and host name both stopped responding at the same time but the ip address ping did not change over to the different address.  From there I found out that the computer I was copying from was the 192.168.1.254 ip address and windows pings it's own ip address\gateway after getting no response from the network for a period of time.  So that part of my post is solved.

 

So now i know that the unraid box is falling off the network in the middle of the transfer. I do not have this problem while playing files back or copying files to other drives.  It only seems to happen on the drive that i freed up space on and then tried to copy files back onto that drive.

 

If the client cannot ping itself it sounds like a local problem. Try a different client computer.

I have tried from a Windows 7 pc, and a macbook pro running the most recent version of OSX.  They both give the same results.  It is the unraid box that is falling off the network in the middle of the transfer not the computer facilitating the transfer.  This only happens when copying the files to disk1.  Never happens during a file read, or when copying,g files to disks that have never had files on them before.

Attach a syslog.

At approximately what time was the failed transfer? The syslog is only 2 minute long. Let the system run for at least 10 minutes, then perform some testing and note the system time of each test. Then post a syslog.

The failure occurred about 5 minutes after the last entry in the syslog.  It did not trigger an event in the syslog for some reason.  I will repeat the failure tonight and let the server sit for 10 minutes after it fails then pull the syslog.  The syslog i posted was pulled around 30 seconds after the file transfer failed and was identical to the one i pulled prior to the error.

Unraid disconnected from the network at 17:25 for around 10 seconds log enough to kill the file transfer.

 

How do you know this?

 

Show the outputs of "ifconfig" and "ethtool eth0".

Two different computer were pinging the unraid boxes ip address and they both stop responding at the time of the failure, and start back up at the same time.

 

I may have inadvertently stumbled into part of my problem.  My disk 4 (not the one that is having issues described in this thread) redballed last night.  I noticed it when I went to pull the event log the second time.  I have a drive on order that will be here in a couple of days.  I'm going to leave my server offline until then, probably preclear the drive on another box, then put it in and let it rebuilt.  I will continue my testing from there.  Hopefully this will fix my problem.  My hopes at this time are that the other disk that was going bad may have been confusing unraid and causing the interruptions.  I'm not very knowledgeable on how the internals of unraid work, but if the software calculating parity was having issues with a drive it seems logical that it could have been freezing up the server.

Were the two different computers pinging each other as well, to rule out network issues?

Not each other, but they were both pinging www.google.com and it never dropped on either of them.

Show the outputs of "ifconfig" and "ethtool eth0".

Show the outputs of "ifconfig" and "ethtool eth0".

 

Attached images of both commands.

ethtool1.pdf

ifconfig1.pdf

OK so after rebuilding my Disk3 last weekend.  It is now having the problem also.  The issue seems to only occur on drives that have had to be replaced and rebuilt from parity.  As of right now I'm only able to copy files to Disk 2.  My Disk1 and my Disk3 both of which have been replaced, give me the file transfer failures.  However, I can copy files to Disk2 at any point and not have any issues.

 

Would upgrading to Unraid 6 Beta 15 be a good idea? It supports an usb3 network adapter i have own so this would allow me to rule out the nic as my issue.

 

Also can i downgrade back to Unraid 5 if it does not help?

So I kept on trying different approaches to this issue and getting nowhere.  The strangest part of it was that I could run a ping on the console and it would never drop, other computers could ping each other and they would not drop either.  Only the ping to the Unraid box from computers would have the drops when this issue surfaced.

 

I upgraded to Unraid 6 Beta 15 and the issue has gone away.  My guess is that it must have been a driver issue that just surfaced for some reason.  My onboard NIC is a Realtek 8111E for reference.  I've been running 5.0.6 since two days after it was released (Oct. 2014), I'm not sure why it just started happening recently (Mar. 2015).  I'm a little worried the onboard NIC on my motherboard in my Unraid server may be getting flakey.  I've transferred 150 gig's worth of data to the server since the upgrade and had zero failures.  So I'm convinced for now that the upgrade has fixed my issue.

 

Thanks guys for all the suggestions and help with this issue.