Ubuntu Networking :: How To Test Cifs Is Actually Mounted In Script File
Mar 12, 2010
I would like to be able to test that a network mounted cifs(samba) share is actually mounted in a script file to do backups. I want to do this so that when my automatic backups run they actually go to the remote location or fail. Currently, if there is a network problem that prevents the network share from mounting, the files simply get copied to the folder (e.g. /media/backupmount) and end up filling up my small local hard drive.
We recently had an issue with "cat /proc/mount" telling us that a CIFS file system was mounted, even though the mount was not working correctly. So we're not sure if we can trust linux to report malfunctioning mounts, so we're planning on adding a specific file on the mounted file system, and verify the mount by reading this file from the client side (linux). If linux fails to read it, we know that the mount have failed. But before we go ahead doing this I thought I'd just hear how others are doing this sort of thing - how do you make sure that mount points are up and working?
- kenneho
EDIT: I just saw that I've posted in the security area, not in the server area. How do I move it?
The problem arises when I try to create a sub-directory inside the mounted directory. All the newly created sub directories become write protected.
I am accessing this file system from R software and it needs to write/create directories in side this mounted directory.
how can newly created sub-directories will become automatically writable, so that R can create new sub-directories and write data inside those directories.
I have a Hitachi SimpleNET adapter (entry-level NAS device) on a Seagate FreeAgent 1TB external HDD (formatted ext3). The NAS device is connected over 100MB/s ethernet to a Netgear Wireless G router. All other devices connect using Wireless G. The NAS runs embedded Linux on an ARM processor and it runs vsftpd and Samba for file transfers.
If I transfer a large file using an FTP client the transfer maxes out at around 2.5MB/s. For my purposes that's good enough, especially considering the Wireless G bottleneck. If I transfer a file from a Windows 7 client (using samba) I get around 2.2MB/s. I know the CIFS protocol has more overhead than FTP and the difference in speed isn't that noticeable.Any combination of Ubuntu and Samba results in me getting less than 1MB/s. I've tried mounting it through Nautilus (GVFS) and /etc/fstab. FTP from this same Ubuntu client gets around 2.5MB/s.
I don't have root access on the SimpleNET to change the smb.conf. I've made a few adjustments to the mount options with no success. how to either speed up 10.04 as a Samba client or mount a folder on an FTP server locally? I've tried both curlftpfs and FUSEFTP. With curlftpfs any write operation results in an I/O error and it crashes intermittently. With FUSEFTP I never got that far and couldn't even browse the folder.
I have mounted a windows network share using the gnome desktop environment, using Places -> Connect To server.The network share is OK, and I have the icon on my desktop and can see all the files.I want to be able to use this network as well in the console, so I need the mount point.What is the location on the filesystem were this networkdrive gets mounted? I find nothing in /mnt and nothing in /media also using mount to look at the registered mounts, there is no entry for the networkdrive.Nevertheless, I have this networkdrive now open in my desktop, and have an option to unmount it.I know that using the mount.cifs command you can specify the mounting point.
When using the following cifs mount command, mount -t smbfs -o username=username,password=password //srv/shr /usr/localfolder/and the cifs share does not exist, localfolder is mounted like d????????? ? ? ? ? ? localfolderafter a number of time , when umounting we get a kern <soft lock>Is there any way to fail the mount if the destination share does not exist, ive had a quick look through man mount but can not see a solution.
I have a line in the fstab file which automatically mounts a network drive every time I start up Ubuntu. I browse to a text file on the network drive and open it using gEdit and make changes to it. Then, when I hit the save button, a bright red warning appears:
Could not save the file [path here] gedit cannot handle file: locations in write mode. check that you typed the location correctly and try again. This also happens if I do save as. Then, after this error appears, the file actually disappears (gets deleted) from the network drive and in order to save it, I have to select save as again and type in the original filename. The line in my fstab file is:
I'm not sure if this has something to do with the file permissions or gEdit itself or using cifs to mount. When I use the "ls -l" command on the file, I get
The shares get mounted correctly and you can navigate through the directories and open files.The only problem is that it randomly starts going really slow taking 30 seconds or longer to open a directory that has 2 or 3 files in it.I have tried quite a few things to try and fix this without any luck. Its getting to the point where I am having to consider recommending that we use windows instead, which I would rather not do as I think its good for students to experience different operating systems during school.
There are a number of shares on the destination system; for the purposes of this thread I used D$ and F$ (corresponding to those partitions). These shares are mounted permanently via CIFS (entries in fstab) on the source system.Today I copied an ISO image of some 3.5 GB from source (S) to destination (D). md5sum on S gave a different checksum for the source ISO than that calculated by HashCheck Shell Extension for the destination ISO. I know some would argue that I shoud use the same md5sum programm for both images.
To circumvent that I 7zipped the ISO, verified it's integrity and copied that archive from S to D. Verification of the acrchive by the Win version of 7z failed.To see if it's a protocol problem I copied both ISO and archive of ISO to another D this time using sshfs (it's an Ubuntu server). Flawless copies.Then I copied both files to another Win-based server on the same network. Flawless copies.Mystified, I checked the partition's file system integrity (NTFS) where the errors occured. Minor inconsistencies (no errors according to chkdsk). So I copied both files again, once to another partition (D:) of the original D, once to that partition causing the error in the first place (F:).
(D:): archive corrupt, checksum okay (F:): this time around both okay.
What the hell can I do to nail down the problem?! I don't even know whether it's a problem of the source system or the destination.
I'm using cifs to mount windows share.I have created one credentials file and given the path in fstab to mount at boot time. Now i want to encrypt the credentials file and place that in the fstab file.But it is not accepting.. how to use encrypted file to use in fstab,so that normal users can not watch the credentials inside the file.
I am trying to install mysql 5.1.44..so i downloaded the binary package, i extracted it and then followed the instructions that were in the manual but i keep getting this error when running this command
Installing MySQL system tables... 100315 20:07:27 [Warning] Can't create test file /var/lib/mysql/mosty.lower-test 100315 20:07:27 [Warning] Can't create test file /var/lib/mysql/mosty.lower-test
I have a little problem with my RHEL5.5 IA64. I mounted a Windows directory with 'mount -t cifs'. After a little while someone else mounted the same windows directory in the same mountpoint. The output from 'mount -v' shows me that the same directory is mounted twice in the same mountpoints. I cannot unmount it, not even with the force option. The error is "Device or resource busy". There are no open files in the shared directory and no one using the directory or subdirectories.
I'm working on testing some software, and I have a question. We have several files of binary data that we need to push through our application to test. It communicates via simple TCP sockets. Is there a way I can send this data to the socket from the command line? I tried doing something like this, but telnet never picked up the data.
if this a simple question I apologise, I'm using a SSH connection to a remote machine which also has ubuntu installed, my remote machine is connected to a windows server, using <places> <Network> and clicking on the server, doing this mounts the server into my remote file systemWhen i look around the file system of the remote machine i'm unable to make the windows server resource available to me.I Assume it has a service file in the /dev directory, but would not know what its called or what i would have to do with it.In the mean time i've managed to connect directily to the server from my local machine, (which is probbialy a better solution) but is it possible to see the server via my remate machine?
Here the description of the issue I am having.I am writing a bash test script which reads lines from a file, builds ISO messages, sends them to a server, reads the response with response code and reports the result of the test to a file or on the screen.The message that I need to send is 94 characters long.Here's the portion of a code that I initially wrote:
#~ Open socket. exec 3<>/dev/tcp/172.26.0.25/9991 #~ Send msg.
I have an NTFS file system nfs-automounted on our RedHat servers. Users can read and write to the file system no problem, and can create new files, edit them, and delete them to their heart's content. The only issue is that utilities such as "dos2unix" cannot create temporary working files:
$ dos2unix events.0818.dat dos2unix: converting file events.0818.dat to UNIX format ... Failed to open output temp file: Operation not permitted dos2unix: problems converting file events.0818.dat
This isn't limited to "dos2unix"; any other utility that creates a temporary working file gets the same problem. If I copy the file to a local file system like /tmp, it works fine. Here's the kicker: this works fine on Solaris systems. I can take the "dos2unix" utility over to a Solaris system that has that exact same NTFS file system automounted via NFS, and it works. No issues creating temporary working files at all.
I have a Western Digital "My Book" on my network which I have mounted with cifs.
If I go into it and vi a file, all is fine. I can write and save and close. When I open the file and add to it and then try to write it again, I get the message:
"thefilename" E212: Can't open file for writing
The file is owned by me still and the permissions are -rw-rw-r--
I don't understand why it works the first time and not the second. Also this same effect is observable when I save from another program to there. The first save is fine, the second can not be saved.
I am experiencing poor performance when using cifs share.Such poor performance occur only in the creation of files on the share but not in the rewrite.This is what i do:
I have been running a server for 3-4 years now, and my shares have been mounting just fine. Well, the network admin looked at a backup and seen that the last date backed up was june. I got to looking around and seen that the share is not mounting. I can mount it with sudo mount -a, which tells me my syntax is correct. I get an error about IPv4 socket not opened and it is aborting the operation when I run dmesg | tail, since I can use the above command to mount later, it sounds to me like it is trying to mount before the network connection is ready.
I have done some looking over some init scripts and found that in the /etc/rc.d/init.r/netfs script it has a line that states that it is checking to see if the network is up before it starts to mount the filesystems and the such. This is set to no, my question is, can I change this option to yes and get my desired results, waiting for the network to be up before it mounts the filesystems.
I'm trying to resume copying from a mounted CIFS device to my local hdd with cURL. I tried
Code: $ curl -C - -O file://myfile and also
Code: $ curl -C <manual offset> -O file://myfile (looked up the manual offset using "$ wc -c")
This resumes copying if I cancel it eg with ^C.
But it does not work if I unmount and remount the CIFS device. cURL then ignores my given offset and continues again from start as if nothing were there without saying a word. With "-C -" the same effect.
I have an ntfs partition that I wish to access as a normal user(non-root). For this I did the following. As root I created a folder /windows and did a chmod 777 -R on /windows. Then I added the following line to /etc/fstab
Now, the partition is mounted alright but the problem is that when any other user (non-root) creates a files in /windows (say by executing touch newfile) the newly created file has the owner and group set as root. The non-root user can create the file and he can also delete the file, however, he cannot change the permissions of the file and also the owner:group is always set as root:root. How do I get across this problem, i.e. how do I mount a partition, so that a non-root user can also change the permissions and ownerships of the files he creates.
I was asking around in some IRC channels earlier trying to develop some thoughts on how NFS is better than CIFS. I set up a FreeNAS file server, and that's where all of my data now resides on a pair of raided drives. That way my main desktop, which is kind of a power hog gaming rig, can be powered off since I pretty much live on my laptop now. Anyway, I began to tinker with CIFS and NFS. Since some family members in the household use Windows, I definitely need CIFS. But I wanted to bounce back to NFS too and check it out.
While I do think it's nice I don't have to worry about authentication to the NAS box when using NFS, it's still a little scary. Being that it's more of a trust method instead of actual authentication, truthfully all a user needs to get into your data is the path to your NFS share and a matching UID. I mean, am I wrong by saying this? Sure, it may seem like NFS is convenient, but this angle of it is a little scary. I just don't feel like that screams "secure."
On the flip side, you have CIFS, which uses a user authentication level. So I hit my little shortcut to my NAS and it asks me who I am. I log in and bam, I have connection. I can browse other folders on the share, etc. This is convenient because I do have a "public" share on here with a generic user. That way if friends come over and want to transfer something to me, I have them drop it in the public share and I later transfer it accordingly. Since there is a user level authentication, this to me seems a little more secure.
Speed wise I was a little concerned, as some users have said NFS is faster than CIFS. Well, they might be right. But I did a few bench tests here on my laptop, using the same exact share except one with CIFS one with NFS. I stayed in the exact same location and transferred the same 300mb file in each instance.
NFS - 1.7mb/s CIFS - 1.5/mb/s
Not exactly enough to warrant a huge argument over, so I leave that argument along the road to be forgot about since it doesn't really have any bearing on this situation. I like things about both NFS and CIFS. I just want to know why is it "not optimal" to use a full blown CIFS setup even if you're using 100% Linux systems.
Install cifs-utils on wheezy armhf apt-get install cifs-utils --fix-missing Reading package lists... Done Building dependency tree Reading state information... Done
using Ubuntu file browser, I browsed my Windows network and logged on to a Windows PC. Now Ubuntu file browser shows me "C$ on WinPC" as a folder. I can open it, read/write files, etc.But from bash prompt, I don't see anything of type CIFS/SMBFS listed in the output of "mount". Only the usual suspects (like local CDROM). How can refer to Windows files from Linux commandline?
I was also having a problem with the Shutdown / Reboot sequence taking ages due to using WiFi, WPA2 and mounting SMB shares. I wasted about 4 hours digging around before I finally realised that the solution involved Upstart. 1. Open a terminal and enter:
Code: sudo gedit /etc/init/network-manager.conf 2. Just below the description line add the following:
Code: pre-stop script /etc/init.d/umountnfs.sh end script
3. Save the script and attempt a restart. I don't know if this will work for everyone, or even what version of Upstart you need for it to work, but it cuts my shutdown time from about 2 mins to about 30 seconds.
Note: This has been working for me about 90% of the time. Occasionally though I see that the script ends prematurely due to the TERM signal and I end up with the 2 minute wait again. I've added the following to the /etc/init/network-manager.conf file:
I'm trying to talk the studio I work at into switching one of the departments to linux. (likely kubuntu). So I'm trialling it, but having issues mounting windows shares.It's working great; all except that only Root can write to the mount. I've tried a few different things with fstab, no go.Below is my fstab so far, and you can see the mountpoints.
Code: # /etc/fstab: static file system information. #
I have a Buffalo Drivestation (model HD-CELU2, 1tb) attached to my network.From my ubuntu desktop I can go to the menu, select "connect to server", put in the ip and share info, and it mounts perfectly.I can open the share and browse eadwrite, but when I try to mount it from a terminal or within fstab, it will still mount, but I cannot see any files that are on the drive. I have about 12gb of data on it, but like I said when I mount it using "mount -t cifs 192.x.x.x/share blah blah blah" I do not see any of the files.If I do a df I can see that the drive has files on it based on the free space available, but if I do an ls nothing shows.
Hopefully this'll be an easy one (but I wasn't able to find any other posts with the exact same problem).I'm connecting to a large hard drive at work. I can mount perfectly fine. The following is the relevant line in my fstab file:
The problem is that when I try to cd to the correct directory, I get a permission denied error. I don't own the mount point, and there aren't general read/write permissions set. But if I change to superuser, I can access it no problem. I can read, write, make directories, etc. So the problem is with my computer--not the remote one.
Now, if I add the option uid=MYID, I can read and write just fine. The system makes me the owner of the directory on mounting. But that's not what I want--I'm trying to allow multiple users access to this file system. I want there to either be a neutral owner (e.g. root) with others having read/write access, or I want the owner of the mount point to be the user currently logged in.
