Windows Users Getting Plagued by Errors

[jwsl224]

Our windows users accessing our TrueNAS build over the SMB are getting plagued by two errors. either "and unexpected network error occurred" (error 0x8007003B) or "handle is invalid". Our TrueNAS build (running on a Dell R730XD) is almost unusable. large file transfer operations need to be baby-sat, and "try again" pressed constantly.

We've tried disable the firewall on the windows machine, disabling the firewall, no joy. All the machines in question are on the same network, so there is no external firewall involved here. they're also 1 or 2 switches from each other. what could be wrong? I understand the OS is free, but it has to be at least usable...

Jared

 

[sretalla]

Let's start with some details of your server...

Hardware description including the NIC and HBA/SATA controller, RAM and CPU. Be specific about the TrueNAS version.

Has it ever worked well? or is this a recent change after an update?

 

[jwsl224]

Let's start with some details of your server...

Hardware description including the NIC and HBA/SATA controller, RAM and CPU. Be specific about the TrueNAS version.

Has it ever worked well? or is this a recent change after an update?

Hi There,
here are the details of the machine:

R730XD (12 x 3.5” + 4 x 3.5” Mid + 2 x 2.5” Rear)
2x Intel Xeon E5-2623v4 2.60Ghz 4 Core Processors
128GB DDR4 PC4-2400T RAM
4x Dell 3.5” 3TB 7.2K 6Gb/s NL-SAS HDDs
HBA330 Pass-Through Controller
Quad Port 1GbE NDC
Rear 2x2.5” Flex Bay Expansion Kit
Mid 4x3.5” Expansion Kit
TrueNAS-12.0-U1.1

and to answer your question, it has never behaved any differently, and neither has our secondary machine running TrueNAS ever behaved differently. It frankly seems like windows and truenas not getting along very well, and was wondering if other people have had similar issues.

Jared

 

[sretalla]

HBA330 Pass-Through Controller
Quad Port 1GbE NDC

I think these 2 are potential sources of the problem.

I would probably start with the NIC. Consider getting an INTEL NIC unless your NIC shows up in ifconfig as igb0-3 or em0-3

I am aware of plenty of folks having no issues with Windows 10 and TrueNAS over SMB. Most if not all of them are using Intel NICs and an LSI HBA.

You may also have some improvement with the current NIC (no guarantee though) by disabling the hardware offload (Network | Interfaces > Edit)

 

[jwsl224]

root@truenas[~]# ifconfig
igb0: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> metric 0 mtu 1500
options=e507bb<RXCSUM,TXCSUM,VLAN_MTU,VLAN_HWTAGGING,JUMBO_MTU,VLAN_HWCSUM,TSO4,TSO6,LRO,VLAN_HWFILTER,VLAN_HWTSO,RXCSUM_IPV6,TXCSUM_IPV6>
ether
inet 192.168.1.20 netmask 0xffffff00 broadcast 192.168.1.254
media: Ethernet autoselect (1000baseT <full-duplex>)
status: active
nd6 options=1<PERFORMNUD>
igb1: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> metric 0 mtu 1500
options=e507bb<RXCSUM,TXCSUM,VLAN_MTU,VLAN_HWTAGGING,JUMBO_MTU,VLAN_HWCSUM,TSO4,TSO6,LRO,VLAN_HWFILTER,VLAN_HWTSO,RXCSUM_IPV6,TXCSUM_IPV6>
ether
inet 192.168.2.20 netmask 0xffffff00 broadcast 192.168.2.254
media: Ethernet autoselect (1000baseT <full-duplex>)
status: active
nd6 options=1<PERFORMNUD>
igb2: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> metric 0 mtu 1500
options=e507bb<RXCSUM,TXCSUM,VLAN_MTU,VLAN_HWTAGGING,JUMBO_MTU,VLAN_HWCSUM,TSO4,TSO6,LRO,VLAN_HWFILTER,VLAN_HWTSO,RXCSUM_IPV6,TXCSUM_IPV6>
ether
media: Ethernet autoselect
status: no carrier
nd6 options=1<PERFORMNUD>
igb3: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> metric 0 mtu 1500
options=e507bb<RXCSUM,TXCSUM,VLAN_MTU,VLAN_HWTAGGING,JUMBO_MTU,VLAN_HWCSUM,TSO4,TSO6,LRO,VLAN_HWFILTER,VLAN_HWTSO,RXCSUM_IPV6,TXCSUM_IPV6>
ether
media: Ethernet autoselect
status: no carrier
nd6 options=1<PERFORMNUD>
lo0: flags=8049<UP,LOOPBACK,RUNNING,MULTICAST> metric 0 mtu 16384
options=680003<RXCSUM,TXCSUM,LINKSTATE,RXCSUM_IPV6,TXCSUM_IPV6>
inet6 ::1 prefixl


hopefully there's nothing in here that'll cause me security issues by posting it :p

 

[HoneyBadger]

The HBA330 is a Dell-branded LSI SAS3008 with preloaded IT firmware; it's the lesser H330 that is a problem and needs the crossflash. Card should be fine.

The (factory default) quad NDC is a Broadcom NetXtreme chipset. Not as good as Intel but definitely not as bad as Realtek. (Edit - Looks like yours has the Intel chipset, did you choose an upgrade?) I would second the suggestion to disable TCP offload, but curious about further info on network setup. @jwsl224 are you using LACP/LAGG or any form of port aggregation on the TrueNAS machine?

 

[jwsl224]

i have not done anything to the network settings. the IP address is handed to the TrueNAS machine by our DHCP server, and i haven't changed anything in it's settings. as far as the switch it is connected to, there is no aggregation configured anywhere on it. and i've had issues with managed and unmanaged switches. are there settings that you recommend i change? also, is TCP offload and Hardware Offload the same thing? also, why is it telling me disabling the hardware offload will decrease performance? isn't that a red flag?

 

[HoneyBadger]

the IP address is handed to the TrueNAS machine by our DHCP server, and i haven't changed anything in it's settings

Generally it's recommended to use a static IP for the TrueNAS machine, but that should only contribute to instability if it was trying to transfer data during a DHCP renewal.

You do have it reachable on two different IPs though (192.168.1.20 and 2.20) - if you're serving SMB on both addresses, and accessing the TrueNAS server by hostname instead of IP, is it possible that your DNS records are somehow flipping between the two?

Also obligatory - check that your cables are seated, in good condition, and you can always replace them proactively because copper is cheap.

are there settings that you recommend i change? also, is TCP offload and Hardware Offload the same thing? also, why is it telling me disabling the hardware offload will decrease performance? isn't that a red flag?

Since you're using Intel NICs and not the Broadcom hardware I was expecting to see, you can likely skip TCP/Hardware offload. It will "decrease performance" in that your CPU would have to perform the TCP encapsulation, but I imagine whatever hardware is in that R730XD is more than up to the task of handling 2Gbps of network traffic.

 

[anodos]

>TrueNAS-12.0-U1.1

Upgrade to latest version.

 

[jwsl224]

Generally it's recommended to use a static IP for the TrueNAS machine, but that should only contribute to instability if it was trying to transfer data during a DHCP renewal.

i don't think that's an issue. but i will make it static, just to rule it out.

You do have it reachable on two different IPs though (192.168.1.20 and 2.20) - if you're serving SMB on both addresses, and accessing the TrueNAS server by hostname instead of IP, is it possible that your DNS records are somehow flipping between the two?

i have it reachable on two IP's, so that computers from the 2.20 network don't have to travel to the router and back to get to the 1.20 network. i'm sure that's not uncommon. but all the computers are accessing the server using IP, not hostname.

Since you're using Intel NICs and not the Broadcom hardware I was expecting to see, you can likely skip TCP/Hardware offload. It will "decrease performance" in that your CPU would have to perform the TCP encapsulation, but I imagine whatever hardware is in that R730XD is more than up to the task of handling 2Gbps of network traffic.

i will try the hardware offloading, because i need to try something. regretfully it doesn't seem that there's anything obviously wrong with the setup. so i'm not sure where to go from here besides limping along as it is.

 

[jwsl224]

>TrueNAS-12.0-U1.1

Upgrade to latest version.

done.

 

[jwsl224]

why does it seem like the interface settings don't stick? when i disable DHCP, disable hardware offloading, and enter a static IP, do a test changes then apply them, everything is back to the screenshot below when i go back to the interfaces page. is that normal?

 

[Patrick M. Hausen]

Networking changes are a three-step process now, so you won't lock yourself out by accident.

1. Apply - saves the changes but does not activate them.
2. Test Changes - activates the changes. If you don't reconnect and save them within 60 seconds, it will revert everything.
3. Reconnect to the UI (if necessary because you changed the address) and Save Changes to make them permanent.

 

[jwsl224]

i understand that. but should the interface settings reflect the changes or does it default to the above view?

Networking changes are a three-step process now, so you won't lock yourself out by accident.

1. Apply - saves the changes but does not activate them.
2. Test Changes - activates the changes. If you don't reconnect and save them within 60 seconds, it will revert everything.
3. Reconnect to the UI (if necessary because you changed the address) and Save Changes to make them permanent.

thank you for that sir. i was not arriving upon the last one.

 

[jwsl224]

ok, the "disable hardware offloading" seems to have only made the problem slightly worse. i need to dig to the bottom of this. how do i get additional error reporting information beyond "and unexpected network error has occurred"? i am using rclone for much of our file transferring needs. are there any flags i can input to extract detailed error information? are there any windows logs i can pull up that would have more information?

right now i'm worried about the integrity of the files on this server. once we fully migrate to it we will have several million files on it. silent file corruption is something that keeps me awake at night..