Window Shares Keep Dropping ( Possible Bug !! )

[MANDINGO]

I have 2 xp computers that i use on the network that i have enabled sharing.  I use these computers for all my .ISO etc and each share uses the pluto structure, now here is the problem the drives seem to drop all the time from the linuxmce system.  I have enable the ping and the disk online check box but the disk online check box dosnt hold after a system reboot etc...  I can see the drives and files if i recheck disk online and do a quick reload then all my movies are back but the system reverts back after awhile.  There are no spaces in my share names, Firewall is disabled on both xp boxes, power managment is disabled on both nics in each of the system, even bought a new gigabit switch so that rules out the physical layer, i am convinced that its a bug  <--- hope not and just something that im doing wrong.  Any help would be apprciated......

[MANDINGO]

Can any1 help me with this problem....  I cant be the only one having issues like this ..........

[agentsmith]

Well, I myself do not have any windows shares (XP) but have Linux shares with samba and do not experience the issue.  I realize this is different from your situation but I wanted to reply so you know your post is getting some attention.

Couple of random thoughts.  So when you say you have eliminated the physical layer, do you mean on all systems?  i.e. have you checked the NIC configuration of your core as well?  Also, while we are on the subject of hardware can you list your specific NIC hardware in your core and the Windowz systems?  Maybe someone here has seen similar issues with that hardware and could provide some insight.  What about firewalls, any on the Windowz systems and how about the configuration on the core?  I assume from your post that the Windowz systems are on the internal NIC of the core?  What router are you using?  Any cables with loose wiring or connectors.  ( I know I know you probably already checked this but want to make sure )

Just a few questions to get some ideas going.

[royw]

Check list:

* using two NICs in your core, one for external LAN, one for internal
* minimize/change connection between core and server. 
* your are using LinuxMCE 0710 RC2 (final release)?  The early betas did have NAS issues.
* try switch protocols between samba and cifs (in pluto-admin, locate the share, then under Filesystem if it has 'cifs', change to 'samba' or vice versa)

Here's more details:  http://wiki.linuxmce.org/index.php/Howto:_Using_Network_Shares

HTH,
Roy

[colinjones]

Quick update to Roy's post - the two parameters for the filesystem box are "cifs" and "smbfs" (not samba).

Also, you can check for that issue by 1) seeing if you can still access the media through KDE desktop /home/public/data/video/windows share <name> or 2) by searching for a piece of media in the UI using the search function and keyboard then playing it - it will work even though the files are not presented in the normal media grid.

Btw the check box is an indicator of status as well as a control and changing of takes effect as soon as you click ok. A reload isn't required. Wait a few seconds after confirming the check box is on and hit refresh, if there is a contectivity issue it will uncheck on it's own.

As Roy said, this was an issue several people were having in previous versions but has been fixed in the RCs so check you are using the latest version

[MANDINGO]

Thanks for the fast responce....  Im using RC2 and i tryed changing from cifs to smbfs and that still didnt help i saw this when i dmesg

[27014.528802]  CIFS VFS: Send error in SessSetup = -5
[27014.657321]  CIFS VFS: cifs_mount failed w/return code = -5
[27018.620834] smb_add_request: request [e9f5af00, mid=6] timed out!
[27042.969475] smb_add_request: request [d355df00, mid=7] timed out!
[27042.969489] smb_lookup: find //.check_mounted failed, error=-5

not sure where to go from here ?

ok so i changed back to cifs and now the dmesg says
[  373.312632]  CIFS VFS: Error 0xffffffed on cifs_get_inode_info in lookup of \.check_mounted
and loops it ?

[colinjones]

hmmm it could be a bug, but proving that and getting it fixed will be an uphill battle - you need to positively eliminate as much as possible first. Get a new/known-good cat5/6 cable and connect one of these machines directly to the core (no switch) - depending on the NICs a normal cable should be fine but if you have a cross-over then use that. This will eliminate your network and switches.

Do you have any spare NIC cards (different brand, preferably one mentioned on the wiki as tested) - try substituting you Internal NIC for this card. Also, please confirm Roy's question on your network setup. You have 2 NICs, one internal one external, and the XP machines are connected to the internal NIC. The internal network has no other DHCP server on it (are there any other devices at all?). The internal and external networks/NICs are not connected to the same switch or any other way to mix the 2 segments?

If no other NICs, then check out your current NIC with the wiki to see if anybody has confirmed compatibility (or at least on a Linux site and look for the latest driver)

[MANDINGO]

My Lmce is a dedicated core not hybrid and uses
Intel Pentium 4 processor (800MHz FSB)
Hyper-Threading enabled
2GB DDRII 400
1 x 64-bit 133MHz Full-height / half length PCI-X Slot
2x Broadcom BCM5721 Gigabit LAN Controllers
Support for IPMI 2.0
One 3.5" IDE 120 Gig drive
1 U case
on the core 1 nic is internal and the other external, the internal is connected to an 8 port D-link gigabit switch and the external is connected to wireless router


Storage box 1 has 2 nics running xp pro, gigabit nic connects to the D-link with a hard coded ip address of 192.168.80.2 and the second 10/100 nic connects to the wireless router hard wired with an ip address of 192.168.1.102 System has 12 500 gig sata drives on a 3ware card with each drive having a windows share like "MOVIES_A-B" with no spaces etc...

Storage box 2 has same setup but gigabit nic has ip address of 192.168.80.3 and the 10/100 nic is 192.168.1.103 also with windows share but sharing out Music

so there is 2 seperate collision domains because of the two different networks and all my MDs connect only to the D-link gigabit switch, This setup allows my regular computers and my wireless laptops to access the storage boxes over the 192.168.1.X network and allows me to dump iso to the drives, while the lmce core has full access to the storage boxes also but over the gigabit network within the 192.168.80.X

Hope this helps but i dont see this being any issue with the errors under dmesg, gonna check on the BCM5721 Gigabit LAN Controllers thats built into the core.  thanks in advance
   

[Zaerc]

What exactly do you mean by "hard coded ip address", and why are you doing that?

[colinjones]

What exactly do you mean by "hard coded ip address", and why are you doing that?

What he said! Why have you specifically created the LMCE internal network exclusively to serve the LMCE environment, but then gone and assigned static IPs - a configuration that has nothing to do with LMCE!

As for the dual-homed XP machines to connect into a wireless network.... why? The stated reason below makes no sense, the second network is completely unnecessary to allow your other PCs access to your storage boxes. Put your wireless directly on your internal or external network and eliminate the dual homing and second network - actually remove the second NICs! Then switch your storage boxes to DHCP.

You never said whether you did the straight cable test I recommended - don't bother until you have straightened out your networking, that adds so much complexity and uncertainty that is unnecessary that you won't be able to draw any conclusions from any testing.

[MANDINGO]

on the internal side of the core it is running dhcp which the mds get there address from and any orbiters but the storage boxes have the ip address assigned by me to 192.168.80.2 and 192.168.80.3 for the second box this should not be a problem just because dhcp is running dosnt mean you cant hard code an ip address to a computer,  After a lease is up on an ip address sometimes a new address can be assigned this is eliminated by me manualy setting the ip

The xp machines have 2 nics for a reason,  first the gigabit network which is the lmce network cannot be seen or touched by any other wireless or wired pc on the 192.168.1.X network.  So theres no need to worry about if the core is translating to clients over the network etc... or some one browsing and deleting files from the core. This allows the core to controll and streams along a dedicated backbone to the house.  The second nic gives me access to the storage boxes over the 192.168.1.X which is where all the everyday computers reside on but never to the 192.168.80.X lmce network allowing me to dump to the storage boxes from any wireless or wired pc on that network which has nothing to do with the lmce network.  Long story short if my 192.168.1.X network goes down the 192.168.80.X network is not effected and still controlls the house, if my core goes down this doesnt effect all the every day computers in the house, I know what i am doing when it comes to networks etc... Its what i do for a living and yes i did try the cable and that is not the issue it still drops and i still get the 

CIFS VFS: cifs_mount failed w/return code = -11


Thanks again

[colinjones]

yes but you don't know about LMCE networks or you wouldn't be asking the question.

Start again with unprejudiced mind - type DHCP into the wiki or forum and read about how LMCE does pnp. It is fundamentally based on DHCP, there are others mechanisms but this is the most fundamental and reliable. The pnp is what adds the media shares correctly so the LMCE and autofs can ensure that they are mounted and available when they are needed. You will probably need to delete the devices and allow it to rediscover them as well.

As a person with significant networking you will be well aware of 1) Windows' long standing troubles with dual-homing, the routeing and publishing of services in that environment, and 2) the need to remove and eliminate any complexity from an environment whilst troubleshooting.

I understand your concern with anything on the core interrupting Internet access but you are mistaken in thinking that your "normal" PCs having access to the LMCE network is a problem. This is desireable! Either directly on the internal network or indirectly routed through the core from the external network - your choice.

Put the wireless device on the external network and add a 445 rule to your core's firewall and all your PCs will have full access to your storage boxes and uninterrupted access to the Internet without the complexity you have. But at least initially remove all that for troubleshooting purposes otherwise I can't help you any further.

[royw]

Take a look at:

  http://forum.linuxmce.org/index.php?topic=4260.msg25037#msg25037

Short summary: try putting the username/password on the share instead of device.

Also you might want to search the forum on the error message.  There are a few hits.

HTH,
Roy

[Zaerc]

on the internal side of the core it is running dhcp which the mds get there address from and any orbiters but the storage boxes have the ip address assigned by me to 192.168.80.2 and 192.168.80.3 for the second box this should not be a problem just because dhcp is running dosnt mean you cant hard code an ip address to a computer,  After a lease is up on an ip address sometimes a new address can be assigned this is eliminated by me manualy setting the ip
...

Actually configuring ip-addresses manually that are within the DHCP range is a huge problem. 

[royw]

on the internal side of the core it is running dhcp which the mds get there address from and any orbiters but the storage boxes have the ip address assigned by me to 192.168.80.2 and 192.168.80.3 for the second box this should not be a problem just because dhcp is running dosnt mean you cant hard code an ip address to a computer,  After a lease is up on an ip address sometimes a new address can be assigned this is eliminated by me manualy setting the ip
...

Actually configuring ip-addresses manually that are within the DHCP range is a huge problem. 

From /etc/dhcp3/dhcpd.conf:

Code: [Select]

subnet 192.168.80.0 netmask 255.255.255.0 {
        next-server 192.168.80.1;
        filename "/tftpboot/pxelinux.0";
        option pxelinux.reboottime = 30;

        default-lease-time 86400;
        max-lease-time 604800;
        pool {
                 allow unknown-clients;
                 range 192.168.80.129 192.168.80.253;
        }
}

The lease pool is .129-.253, so it's ok to have static IP machines in .2-.128.  I've had no problems having a static IP server (.20) with a NFS share.  Simply follow the instructions on the wiki to manually add the servers and shares.

One note in favour of using DHCP is that once LinuxMCE assigns an IP to a file server, it binds the MAC address to a fixed IP address (see the group section in dhcpd.conf), so for all practicable purposes the server's address is static (at least until you change NIC cards).

I suspect that Collin is on the right track suspecting that dual homing the servers is the culprit.  It might be time to dust off wireshark to see exactly why the connections to the CIFS shares are being rejected.

HTH,
Roy