Grime121
asked on
MrxSmb 3019 Redirector Failed to Determine Connection Type
When copying a file or files to or from a shared folder on either one of my w2k3 servers I usually get this warning in the event log:
Event Type: Warning
Event Source: MRxSmb
Event Category: None
Event ID: 3019
Description: The redirector failed to determine the connection type.
Data:
0000: 00 00 00 00 04 00 4e 00 ......N.
0008: 00 00 00 00 cb 0b 00 80 ....Ë..€
0010: 00 00 00 00 84 01 00 c0 ....„..À
0018: 00 00 00 00 00 00 00 00 ........
0020: 00 00 00 00 00 00 00 00 ........
I've read almost everything there is to find on this error. The shared folders I am copying to and from aren't mapped, so it's not a problem with 'net use'. It's not a problem with a poor connection because there is not much traffic and the servers are plugged into the same switch. I don't think it's a driver issue because it's happening on two different servers with two completely different NIC cards (plus I've tried a second NIC card in one of the servers). I know one of Mircosoft's KB articles says that "They have confirmed it is a problem" and "This can be safely ignored", but it's happening almost every time, and it's cluttering my logs. If indeed these errors are only being caused by a flaw in the operating system, is there really no hotfix for this? Does anyone know of anything else I can try?
Event Type: Warning
Event Source: MRxSmb
Event Category: None
Event ID: 3019
Description: The redirector failed to determine the connection type.
Data:
0000: 00 00 00 00 04 00 4e 00 ......N.
0008: 00 00 00 00 cb 0b 00 80 ....Ë..€
0010: 00 00 00 00 84 01 00 c0 ....„..À
0018: 00 00 00 00 00 00 00 00 ........
0020: 00 00 00 00 00 00 00 00 ........
I've read almost everything there is to find on this error. The shared folders I am copying to and from aren't mapped, so it's not a problem with 'net use'. It's not a problem with a poor connection because there is not much traffic and the servers are plugged into the same switch. I don't think it's a driver issue because it's happening on two different servers with two completely different NIC cards (plus I've tried a second NIC card in one of the servers). I know one of Mircosoft's KB articles says that "They have confirmed it is a problem" and "This can be safely ignored", but it's happening almost every time, and it's cluttering my logs. If indeed these errors are only being caused by a flaw in the operating system, is there really no hotfix for this? Does anyone know of anything else I can try?
RWrigley
Is your server multi-homed?
ASKER
one of the servers has VPN enabled using Routing and Remote Access. It has one IP for the RRAS connections, and one for the LAN. I seem to remember these warnings beginning to show up when I installed RRAS too.... The other server is the file server though, and so these warnings appear on it a lot more often. It does not have RRAS enabled, and only has one IP.
have you tried increasing the auto-disconnect setting:
HKLM\System\CurrentControl
The value is in minutes, so put it something huge.
HKLM\System\CurrentControl
The value is in minutes, so put it something huge.
ASKER
I just tried that, and it didn't seem to help. It happens especially when copying files from a folder share on one computer to a folder share on another. The error appears immediately after I start copying or moving the files from one share to another.
Hmmm...what model of NIC, and what hubs/switches between the comptuers?
ASKER
One of the servers has a Realtek integrated gigabit card, and the other has an integrated intel pro 1000 XT card. Latest drivers on both of them.
Also, I'm raising the point value to 200.
Also, I'm raising the point value to 200.
What kind(s) of hub/switch(es) are between them?
ASKER
Oh yeah, sorry. There are actually two switches between them. Both switches are linksys. One switch is a 5-port linksys that also has the main router, and several other switches plugged into it. One of the switches plugged into that one is a 16-port linksys switch that has the other server plugged in to it. The response time when I ping it is <1 and there is no packet loss.
I'd be more concerned about sustained throughput then the ability to pass through a single packet. :-)
Can I assume that neither of the switches supports Gigabit? If so, you may be experiencing a reset cuased by you network cards trying to send data faster then the hardware can accept it. If so, you can try configuring the network cards to 100tx, instead of autosensing (which is the default).
Can I assume that neither of the switches supports Gigabit? If so, you may be experiencing a reset cuased by you network cards trying to send data faster then the hardware can accept it. If so, you can try configuring the network cards to 100tx, instead of autosensing (which is the default).
ASKER
Yeah, I've actually tried that as well. I also tried turning on the "Optimal Performance" setting on the Realtek card. None of that helped.
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER
Well, one server is a PowerEdge 2600 that I got from Dell less than a month ago. I've checked their website since then and there have been no driver updates for it. The other is a server that was just sort of thrown together. It has a Giga-byte motherboard. The driver I have I got from their site http://www.giga-byte.com/Motherboard/Support/Driver/Driver_GA-7VAXP.htm . It doesn't say specificly that it works with w2k3. It works with all other versions of Windows though....
As for the network cables, the cable to the PE 2600 is new. The cable to the other server is not. I will replace it when I get back to the office and see if that helps.
As for the network cables, the cable to the PE 2600 is new. The cable to the other server is not. I will replace it when I get back to the office and see if that helps.
Hi Grime121...
Did you find a solution? I have the same problem...
Regards
Morten
Did you find a solution? I have the same problem...
Regards
Morten
ASKER
Nope. It's been a year now and I still haven't found a solution. I pretty much gave up on this one.
When you get this error are the performance slow as well? and where are you copying to (Windows XP, Windows terminal server or???)
ASKER
Doesn't seem to matter whether I'm copying to another server (w2k3) or a workstation (winxp). It even happens when I download stuff from the ftp server sometimes if the server has to pull the data i'm downloading from another server. I think one thing in particular that causes the error is if you transfer a file from the server to a network share on the same server. For instance, if you have a file sitting on the desktop and instead of copying it to D:\FTP_FILES you copy it to \\SERVER\FTP_FILES. It has something to do with using the loopback address.
From the things that I've read about this it is a common problem that does not seem to have a solution. It does not do any damage as far as I'm aware though. I have not seen any noticable performance decreases.
From the things that I've read about this it is a common problem that does not seem to have a solution. It does not do any damage as far as I'm aware though. I have not seen any noticable performance decreases.
ASKER
I think I need to elaborate... The example that I gave is really what causes the error for me more than anything else. I have made it a point to stop transfering files in this manner, and I now receive far less of these errors.
Note,
Even thought this is from Dec 2007 MS it may help :
SYMPTOMS
When you map a network drive to a local share, the following error message may be reported in the system event log:
Event ID: 3019
Source: MRxSmb
Description: The redirector failed to determine the connection type.
Back to the top
CAUSE
This behavior can occur when NetBIOS over TCP/IP (NetBT) attempts to query the destination device (in this case, the loopback adapter) for network speed. The loopback adapter does not handle speed negotiation and therefore cannot negotiate the speed.
This behavior is associated only with the TCP/IP protocol because TCP/IP is the only protocol that uses the Loopback adapter.
Even thought this is from Dec 2007 MS it may help :
SYMPTOMS
When you map a network drive to a local share, the following error message may be reported in the system event log:
Event ID: 3019
Source: MRxSmb
Description: The redirector failed to determine the connection type.
Back to the top
CAUSE
This behavior can occur when NetBIOS over TCP/IP (NetBT) attempts to query the destination device (in this case, the loopback adapter) for network speed. The loopback adapter does not handle speed negotiation and therefore cannot negotiate the speed.
This behavior is associated only with the TCP/IP protocol because TCP/IP is the only protocol that uses the Loopback adapter.
I realize this question is already closed, but just to help show another correlation, I'd be interested to see how many people finding this question have ESET's NOD32 anti-virus program (version 3) installed on either their client or the server.
Note that 3019 is a warning, but it does cause a slowdown which you may or may not notice. There is overhead associated with file operations. If you are seeing this error, you MAY have very slow copies over the network. You'd probably only notice this if you were copying thousands of TINY (< 1k) files. You will probably not notice it if you are copying medium to large files, as the throughput remains high with this warning, but the overhead at the beginning of each file takes longer.
Note that 3019 is a warning, but it does cause a slowdown which you may or may not notice. There is overhead associated with file operations. If you are seeing this error, you MAY have very slow copies over the network. You'd probably only notice this if you were copying thousands of TINY (< 1k) files. You will probably not notice it if you are copying medium to large files, as the throughput remains high with this warning, but the overhead at the beginning of each file takes longer.
Does anyone know the reg keys to disable the logging of MrxSMB altogether? I don't want warnings, I don't want Errors, and I don't want informational. It is a performance hit that is unjustified, and unprofessional on Microsoft's part.
Hi,
Were also getting this and it's causing severe network problems. Everything seemed to be fine until we moved the file share from a workstation to the SBS2003 server. Since then there are a mixture of network problems, as listed:
" Files are slow to open.
" Some clients are unable to print but all can run a test print.
" Machine performance is slower.
We also have ESETs NOD32 AV version4 installed on clients and server.
Ive tried:
" Removing Internet Explorer Enhanced Security Configuration from server add/remove programs.
" Set reg key HKLM\System\CurrentControl
" Turned off local device mapping for printer etc when using Remote Desktop.
" Resynchronised the time server using command w32tm /s on one machine which had time errors and its battery replaced.
All seemed to have temporally effects but then the problem continued.
The Server is running at 1Gb NIC connected to a 1Gb switch which is then connected to the LAN clients via a 100Mb switch. There is a WLAN connected the 1Gb switch also. Both wired and wireless machines seem to be effected. The two switches are connected via a Cat5e cable using the Uplink port on the 100Mb switch.
( As this is closed I will post the question to another posting)
Were also getting this and it's causing severe network problems. Everything seemed to be fine until we moved the file share from a workstation to the SBS2003 server. Since then there are a mixture of network problems, as listed:
" Files are slow to open.
" Some clients are unable to print but all can run a test print.
" Machine performance is slower.
We also have ESETs NOD32 AV version4 installed on clients and server.
Ive tried:
" Removing Internet Explorer Enhanced Security Configuration from server add/remove programs.
" Set reg key HKLM\System\CurrentControl
" Turned off local device mapping for printer etc when using Remote Desktop.
" Resynchronised the time server using command w32tm /s on one machine which had time errors and its battery replaced.
All seemed to have temporally effects but then the problem continued.
The Server is running at 1Gb NIC connected to a 1Gb switch which is then connected to the LAN clients via a 100Mb switch. There is a WLAN connected the 1Gb switch also. Both wired and wireless machines seem to be effected. The two switches are connected via a Cat5e cable using the Uplink port on the 100Mb switch.
( As this is closed I will post the question to another posting)
I am also getting this event id:3019 on my server. My workstations will try to access the mapped drives and nothing will be there in the browse window. I have ESET Nod32, I have not moved to the version 4. I restart the server and the workstations and everything connects. This is a frustrating issue. Nod32 is pretty good with virus protection from my experience with it.
wjl3698, if I remember right, ESET fixed a lot of their server issues with a later 3.x version of their antivirus. Search their forums for 3019 warnings and 6004 errors to find the exact version -- or just upgrade to the latest 3.x version on your server side.
PS: In relation to my earlier comment about many tiny files. Microsoft has actually documented that issue in KB 321098: http://support.microsoft.com/kb/321098
I'm not sure why it only seemed to show up with NOD32 installed, but I was able to demote that particular server from being a domain controller and haven't had issues since.
I'm not sure why it only seemed to show up with NOD32 installed, but I was able to demote that particular server from being a domain controller and haven't had issues since.
So far none of the fixes have worked for us. We have event tried disabling most periodic WAN traffic in addition. http://support.microsoft.com/kb/819108. We are looking for a way to prevent the logging of MrxSmb warnings and errors altogether via registry without disabling the event log service altogether,
I had same problem.
My error popped up in the system log whenever I would edit a login script on 1 AD controller.
The changes would replicate immediately to my 2nd DC but that 3019 event ID and The redirector failed to determine the connection type.
All I did was go to IE in both controllers.
Go to tools, options, security, local intranet, sites and add both DC's in this format.
\\server01
\\server02
After you apply and save, it will change the formatting to file://server01 and file://server02
The error never came back.
My error popped up in the system log whenever I would edit a login script on 1 AD controller.
The changes would replicate immediately to my 2nd DC but that 3019 event ID and The redirector failed to determine the connection type.
All I did was go to IE in both controllers.
Go to tools, options, security, local intranet, sites and add both DC's in this format.
\\server01
\\server02
After you apply and save, it will change the formatting to file://server01 and file://server02
The error never came back.
I missed this question because it wasn't within the Netbios zone:
The network Redirector is used for netbios translation. The redirector has different NODE types. Sometimes, you can tell DNS to attempt to provide Netbios translation. That can confuse the redirector because it doesn't understand the traditional node type.
I missed this post. If anyone needs help with this, I would like to see an open question to help them overcome this issue.
For now, let's do a little troubleshooting.
At the command prompt type:
Net config redir
There you should see SMB 0000000000 and just below that a MAC address for Netbios translation. NOW, type IPconfig /all at the command prompt and pay attention to the node type.
Review the current node type and make sure that your MAC address is appropriate with the server's MAC. If more than one MAC then Netbios is bound to two or mor nic cards. This is a problem for netbios.
The reason Microsoft says it can safely be ignored is because they are trying to go to full DNS. But Netbios offers some services that DNS can't quite accomodate.
The network Redirector is used for netbios translation. The redirector has different NODE types. Sometimes, you can tell DNS to attempt to provide Netbios translation. That can confuse the redirector because it doesn't understand the traditional node type.
I missed this post. If anyone needs help with this, I would like to see an open question to help them overcome this issue.
For now, let's do a little troubleshooting.
At the command prompt type:
Net config redir
There you should see SMB 0000000000 and just below that a MAC address for Netbios translation. NOW, type IPconfig /all at the command prompt and pay attention to the node type.
Review the current node type and make sure that your MAC address is appropriate with the server's MAC. If more than one MAC then Netbios is bound to two or mor nic cards. This is a problem for netbios.
The reason Microsoft says it can safely be ignored is because they are trying to go to full DNS. But Netbios offers some services that DNS can't quite accomodate.
Hello,
We are having the same/similar problem. Event viewer is reporting the same, but symptoms are a bit different.
My environment is WIN2K3 server with Wyse thin clients connected via RDP. I have a total of about 30 clients connected but only a handful of clients continuously freeze and have to restart their sessions to re-connect. It does not happen everyday. In Task Manager, I will see them connect and get bumped right out. I have yet to find a common denominator.
Thanks for any prompt feedback.
We are having the same/similar problem. Event viewer is reporting the same, but symptoms are a bit different.
My environment is WIN2K3 server with Wyse thin clients connected via RDP. I have a total of about 30 clients connected but only a handful of clients continuously freeze and have to restart their sessions to re-connect. It does not happen everyday. In Task Manager, I will see them connect and get bumped right out. I have yet to find a common denominator.
Thanks for any prompt feedback.
I have the same problem, 2003 server it happens when I try to connect with logmein, no exact time frame, some connections are maintained for 2 - 30, then I might a couple on minutes, then back to seconds; no time pattern to getting kicked out.