Buy Needles And Syringes With No Prescription
M4B Store Banner
intex
Riptropin Store banner
Generation X Bodybuilding Forum
Buy Needles And Syringes With No Prescription
Buy Needles And Syringes With No Prescription
Mysupps Store Banner
IP Gear Store Banner
PM-Ace-Labs
Ganabol Store Banner
Spend $100 and get bonus needles free at sterile syringes
Professional Muscle Store open now
sunrise2
PHARMAHGH1
kinglab
ganabol2
Professional Muscle Store open now
over 5000 supplements on sale at professional muscle store
boslabs1
granabolic1
napsgear-210x65
monster210x65
over 5000 supplements on sale at professional muscle store
over 5000 supplements on sale at professional muscle store
DeFiant
UGFREAK-banner-PM
STADAPM
yms-GIF-210x65-SB
over 5000 supplements on sale at professional muscle store
over 5000 supplements on sale at professional muscle store
wuhan2
dpharma
marathon
zzsttmy
over 5000 supplements on sale at professional muscle store
over 5000 supplements on sale at professional muscle store
azteca
crewguru
advertise1x
advertise1x
over 5000 supplements on sale at professional muscle store
over 5000 supplements on sale at professional muscle store
over 5000 supplements on sale at professional muscle store
over 5000 supplements on sale at professional muscle store
over 5000 supplements on sale at professional muscle store
over 5000 supplements on sale at professional muscle store
over 5000 supplements on sale at professional muscle store

User home directory doesn't map at random times

mooshue

New member
Registered
Joined
Apr 3, 2008
Messages
298
Usually i can pinpoint the root of an issue, i like to do that so i can prevent it from happening again on my network but this one has me stumped so i hope kaiser can give some input.

Environment with the issue is as follows on a VMWARE ESX 3.5.0 server

Data eblas-dc1 - Server 2003 SP2
Domain Controller/DNS - Server 2003 SP2
Exchange 07 - Server 2003 SP2
Client PC's - (5 clients on Vista Business SP1) rest on XP Pro SP3


Users share is locaed on the data server UNC \\eblas-dc1\users$\

Users report that their user directory (U:\) does not map for them when they log in. Some users report that the drive (U:\) does map, but not directly to %username% like it is supposed to, but in the root of the shared directory called USERS\ which lets them see everyone else's user folder.

I had the user's home directory set in the active directory profile to like so
\\eblas-dc1\users$\%username%

Im able to map it manually without issue through script and cmd prompt, but for some reason i can not pinpoint it sometimes for random users will not map on login like its supposed to.

If i can not find the root cause to prevent it, i will have to manually set permissions on each user's directory, this may be good practice anyway.
 
Last edited:
Set the following in Group Policy above the WORKSTATION OU:

Computer Configuration\Admin Templates\System\Logon
Always Wait For Network=Disabled

Then this above the USER OU:

User Config\Admin Templates\System\User Profile
Connect Home Directory To Root Of Share=Disabled

As for the problem of not mapping to the %username%, that sounds like a deep share issue and the above might help. By deep share I mean Microsoft just allowed this late in NT4 days as Novell had it for years prior. Have you tried using the following syntax:

\\eblas-dc1.domain.local\users$\%username%, subsituting .domain.local with your root, of course?

What could this will essentially take NetBIOS out of the question. For instance if you have a non-DC server that is multi-homed, such as 2 NICs pointing to different subnets, it senses that on the subnet that is not associated with the DCs subnet does not have a master browser so it becomes master browser. Once a node is a master browser, it is a master for ALL subnets it is tied to. Therefor it will try to be a master on the subnet that the DC is on. The DC however knows that it should take precidence, so it doesn't reliquish master either. You end up having what we call NetBIOS browsing wars where all of a sudden all nodes can only see themselves, believe it or not. Well, fully qualifying it will make it so that the workstations uses DNS instead of NetBIOS so it will force the proper resolution of the server IP where the share is located. Example: some people multi-home Exchange servers for obvious reasons.

NOTE: The previous can also happen if a user brings a Unix or Linux box on the subnet and has SAMBA (SMB=server message block) enabled.

Check your DC for Kerberos errors. This may be caused by Vista or XP firewalll rules blocking the route. You can remediate this by turing the GP off at the offending workstaions and diagnosing it, OR setting Group Policy to turn off Firewalls completely over your OU.

Computer Configuration\Admin Templates\Network\Network Connections\Windows Firewall\Domain AND Standard Profile

Protect All Network Connections=Disabled

(You can obviously change this back after diagnosis, but do not do this on laptop OU as if you are still diagnosing when a user leaves, they will be exposed on external networks)

If you do have Kerberos or Authentication errors please list them out for me, also look on the workstations themselves.

Are any of the workstations going through a firewall appliance to get to the DCs? If so, this may be your problem also. Set GP to do a TCP/IP handshake on authentication instead of running it over UDP. With certain firewalls even firewalls that are third party ON the workstations or servers themselves like Norton, etc, will sense UDP kerberos authentication as an SPI firewall exploit.
Needless to say make sure that ALL firewalls, software or hardware on your internal LAN are disabled while diagnosing. To set the registry for this change, run a REG file that you create with the following KB article then reboot:

How to force Kerberos to use TCP instead of UDP in Windows

Next, lets check out your DNS. How is it's integrity? I am assuming you are running Microsoft's DNS instead of a third party, so check your event log on your DCs. Of course ignore DNS errors while a server is starting up because the caveat is that Kerberos Authentication of the DNS service cannot autheticate BECAUSE it has no DNS serrvice running at the time. It's a chicken before the egg symtpom. Once it is started, DNS can finally allow for Kerberos and the error will stop occuring. I know, it's dumb, but it only happens on startup for that reason. Once started, and all DCs are started do a issue a net stop netbios and then a net start netbios. This will make sure that all proper Active Directory DNS records are completely registered. Go into DNS managment itself and make sure there are no wrong server DNS records. Confirm that you have a reverse DNS zone properly configured. If you dont make sure you create a reverse on every DNS server you have, as Kerberos does a reverse lookup to authenticate. If you don't the workstation may not be able to establish it's own secure channel to the DCs or other servers. This will prohibit the User themselves from authenticating correctly as well. By default, you authenticate by Kerberos, then NetBIOS if Kerberos fails. UNLESS you have GP locked down to only allow for Kerberos.

As you have Exchange on your network, you more than likely still have the Support Tools instealled on your DCs as DCDiag is a prescribed step in setting up Exchange. Run dcdiag > c:\dcdiag.txt. Then post it here if there are errors and I will tell you how to remediate them.

Once you issue new GP settings above, make sure you reboot a couple times, OR issue a gpupdate prior to rebooting the first time. If you still have a problem after that, I want you to confirm that the GPs are active by running rsop.msc. Check for the active settings.

I'm sure you probably did, but sometimes, even when issued through ADUC, the original mapped drive will not give up it's mount for a new mapped destination. So essentially, if a user is already mapped to the root of the share instead of the deep share, even though ADUC has a user profile mapped to it's deep share it will not map it because the workstation itself is hanging on to the old mapping. To remediate this temporarily as it will only take one time to fix, is to issue a "net use u: /delete" (assuming "U: is their %homedrive%. They will just need to log off, then back on two times.

Now we move on to permissions. Make sure that the users you are talking about do indeed have NTFS permissions and correct share permissions. IF your original admin knew what he was doing he would normally lock out users from the root so they cant browse the root directory. For this however, the default of "Bypass Traverse Checking". What this does is, it essnetially makes it so that NTFS bypasses all root folder permission checking and only checks the final destination folder to see if te user has access. If this setting is DISABLED, you will NOT map correctly in this situation.

Let me know what you come up with.
 
BTW, I mentioned Group Policy over an OU. This will ensure that local GP does not override it. You will be able to confirm that with rsop.msc, to see which is active. You are of course able to set local GP on an offending machine instead, but I like Domain GP over an OU for simplistic reasons.

If you are unsure of the setting I mentioned, create a temp OU and move a couple of offending test nodes to that OU and offending users to that OU as well to test the GP only over that particular temp OU.
 
thanks man, very thorough. Im going to start with the first fix as well as skim the log for kerbos errors and see if that will narrow anything down. I was really not sure where to start but the netbios and authentication errors make sense.
 
ill keep responses in public just in case anyone else ever tries to search this issue, hopefully it will turn up with some good SEO.

Kaiser, See comments in red below within your post, again thank you for all of the great info, i kept digging further and further and fixed a lot of little issues i found along the way.



Original post
Set the following in Group Policy above the WORKSTATION OU:

Computer Configuration\Admin Templates\System\Logon
Always Wait For Network=Disabled

Then this above the USER OU:

User Config\Admin Templates\System\User Profile
Connect Home Directory To Root Of Share=Disabled

This fixed the root share issue

As for the problem of not mapping to the %username%, that sounds like a deep share issue and the above might help. By deep share I mean Microsoft just allowed this late in NT4 days as Novell had it for years prior. Have you tried using the following syntax:

\\eblas-dc1.domain.local\users$\%username%, subsituting .domain.local with your root, of course?

Changed the user home path to \\eblas-dc1.domainname.local\users$\

What could this will essentially take NetBIOS out of the question. For instance if you have a non-DC server that is multi-homed, such as 2 NICs pointing to different subnets, it senses that on the subnet that is not associated with the DCs subnet does not have a master browser so it becomes master browser. Once a node is a master browser, it is a master for ALL subnets it is tied to. Therefor it will try to be a master on the subnet that the DC is on. The DC however knows that it should take precidence, so it doesn't reliquish master either. You end up having what we call NetBIOS browsing wars where all of a sudden all nodes can only see themselves, believe it or not. Well, fully qualifying it will make it so that the workstations uses DNS instead of NetBIOS so it will force the proper resolution of the server IP where the share is located. Example: some people multi-home Exchange servers for obvious reasons.

NOTE: The previous can also happen if a user brings a Unix or Linux box on the subnet and has SAMBA (SMB=server message block) enabled.

Good to know, I have a 1TB machine I was about to load as a NAS with SAMBA


Check your DC for Kerberos errors. This may be caused by Vista or XP firewalll rules blocking the route. You can remediate this by turing the GP off at the offending workstaions and diagnosing it, OR setting Group Policy to turn off Firewalls completely over your OU.

Computer Configuration\Admin Templates\Network\Network Connections\Windows Firewall\Domain AND Standard Profile

Firewall was already set to be disabled in GP

Protect All Network Connections=Disabled

(You can obviously change this back after diagnosis, but do not do this on laptop OU as if you are still diagnosing when a user leaves, they will be exposed on external networks)

If you do have Kerberos or Authentication errors please list them out for me, also look on the workstations themselves.

No Kerberos or authentication errors on any servers logs

Are any of the workstations going through a firewall appliance to get to the DCs? If so, this may be your problem also. Set GP to do a TCP/IP handshake on authentication instead of running it over UDP. With certain firewalls even firewalls that are third party ON the workstations or servers themselves like Norton, etc, will sense UDP kerberos authentication as an SPI firewall exploit.

I have a sonicwall tz170 as my gateway/wireless access point but lan settings are set to allow all traffic within lan. Will monitor log to see if any conflicts on lan traffic

Needless to say make sure that ALL firewalls, software or hardware on your internal LAN are disabled while diagnosing. To set the registry for this change, run a REG file that you create with the following KB article then reboot:

How to force Kerberos to use TCP instead of UDP in Windows

Next, lets check out your DNS. How is it's integrity? I am assuming you are running Microsoft's DNS instead of a third party, so check your event log on your DCs. Of course ignore DNS errors while a server is starting up because the caveat is that Kerberos Authentication of the DNS service cannot autheticate BECAUSE it has no DNS serrvice running at the time. It's a chicken before the egg symtpom. Once it is started, DNS can finally allow for Kerberos and the error will stop occuring. I know, it's dumb, but it only happens on startup for that reason. Once started, and all DCs are started do a issue a net stop netbios and then a net start netbios. This will make sure that all proper Active Directory DNS records are completely registered. Go into DNS managment itself and make sure there are no wrong server DNS records. Confirm that you have a reverse DNS zone properly configured. If you dont make sure you create a reverse on every DNS server you have, as Kerberos does a reverse lookup to authenticate. If you don't the workstation may not be able to establish it's own secure channel to the DCs or other servers. This will prohibit the User themselves from authenticating correctly as well. By default, you authenticate by Kerberos, then NetBIOS if Kerberos fails. UNLESS you have GP locked down to only allow for Kerberos.

As you have Exchange on your network, you more than likely still have the Support Tools instealled on your DCs as DCDiag is a prescribed step in setting up Exchange. Run dcdiag > c:\dcdiag.txt. Then post it here if there are errors and I will tell you how to remediate them.

DCdiag passed every test
Browsed through all dns entries, saw a few that never pruned from old machine setups and old backup server ip’s. Resolved all of those.
Reverse lookup was a mess, about 100 duplicate ip’s, wrong backup server ip’s to other subnets. Etc.. resolved all of those


Once you issue new GP settings above, make sure you reboot a couple times, OR issue a gpupdate prior to rebooting the first time. If you still have a problem after that, I want you to confirm that the GPs are active by running rsop.msc. Check for the active settings.

resultant set of policy confirmed changes were applied with gpupdate and reboot on dc

I'm sure you probably did, but sometimes, even when issued through ADUC, the original mapped drive will not give up it's mount for a new mapped destination. So essentially, if a user is already mapped to the root of the share instead of the deep share, even though ADUC has a user profile mapped to it's deep share it will not map it because the workstation itself is hanging on to the old mapping. To remediate this temporarily as it will only take one time to fix, is to issue a "net use u: /delete" (assuming "U: is their %homedrive%. They will just need to log off, then back on two times.

Added net use u: /delete as well as all other network drives to beginning of login script to prevent any further mapping issues.

Now we move on to permissions. Make sure that the users you are talking about do indeed have NTFS permissions and correct share permissions. IF your original admin knew what he was doing he would normally lock out users from the root so they cant browse the root directory. For this however, the default of "Bypass Traverse Checking". What this does is, it essnetially makes it so that NTFS bypasses all root folder permission checking and only checks the final destination folder to see if te user has access. If this setting is DISABLED, you will NOT map correctly in this situation.

Let me know what you come up with.
 
So are your problems resolved?
 
Also, keep in mind, the net use /delete should be a one time thing with U drive as the drive gets instantiated prior to logon script. The was meant to be a cleanup only, so remove it once problems resolved as you will be essentially removing the correct mapped drive after they have already succesfully mapped to the correct path. In other words, if your problem is resolved, remove it from your logon script.

You can keep it for all other drives in logon script, just not drives that are mapped via GP, ADUC. or software distribution points. I say software distribution points because an install may require a reboot, then to continue on to install software. During login if it disconnects a drive prior to the installer making a call to that drive to conitnue an installation, the installation will essentially fail, and most of the time without warning. But the install will be faulty if not totally corrupt. You can assure this also by setting OU GP over all nodes to run logon scripts in sync instead of async.
 
Last edited:
Sorry for separate posts for resolve...

As for DNS, make sure you set the pruning age for retired DNS entries under properties of BOTH reverse and forward zones. To do this right click the subnet in DNS or zone in DNS and go to properties. Click "Aging", then check the box at the top that says "Scavenge Stale Resource Records". Leave the default of 7 days unless your DHCP is not default 7 day lease.
 
The problem of mapping to the root of the users share is resolved.

I do not yet know if all the users problems are resolved of the drive just plain not mapping at all yet as it has only been 1 day, and the issue seemed to happen at random with random people. I will give it a few days and do a check with all of those users that reported the issue.

I am contemplating just setting the GP to use the local MY DOCUMENTS folder as the shortcut to the users home directory. Do you have any opinion on this good bad or otherwise?

I just figured this would eliminate the need for the mapping all together as it would just make the necessary registry changes on the local machine with an authenticated domain profile.

This would also clear up user ignorance. I stress time and time again as im sure you know, do NOT save shit on your computer, use the network drives and place things where they belong so i can restore from backup if needed. Users either just don't get it or just can't be bothered with saving things properly. Since they seem hell bent on ignoring me and using the My document folder anyway, even though i change the user settings in the shell registry keys to use the home directory as the default directory, this would solve this issue as well.

Just a side industry question, do you see that users tend not to report issues and just deal with them for long periods of time, then get all pissy because this problem that they have never reported has been happening for a year and they need it fixed yesterday? lol
 

Staff online

  • LATS
    Moderator / FOUNDING Member / NPC Judge

Forum statistics

Total page views
576,134,870
Threads
138,450
Messages
2,857,165
Members
161,444
Latest member
asd222
NapsGear
HGH Power Store email banner
yourdailyvitamins
Prowrist straps store banner
yourrawmaterials
3
raws
Savage Labs Store email
Syntherol Site Enhancing Oil Synthol
aqpharma
yms-GIF-210x131-Banne-B
hulabs
ezgif-com-resize-2-1
MA Research Chem store banner
MA Supps Store Banner
volartek
Keytech banner
thc
Godbullraw-bottom-banner
Injection Instructions for beginners
YMS-210x131-V02
Back
Top