I am trying to do an rsync that includes a symlink in a folder. I only want the symlink to change on the destination if the mod time of the symlink on the source is newer. The behavior i am seeing is that the symlink on the destination is always updated to match the source regardless of the mod time.
View 2 RepliesThought I'd post it here because it's more server related than desktop... I have a script that does:
[Code]....
This is used to sync my local development snapshot with the live web server. There has to be a more compact way of doing this? Can I combine some of the rsyncs? Can I make the rsync set or keep the user and group affiliations? Can I exclude .* yet include .htaccess?
When I run rsync --recursive --times --perms --links --delete --exclude-from='Documents/exclude.txt' ./ /media/myusb/
where Documents/exclude.txt is
- /Downloads/
- /Desktop/books/
the files in those directories are still copied onto my USB.
And...
I used fetchmail to download all my gmail emails. When I run rsync -ar --exclude-from='/home/xtheunknown0/Documents/exclude.txt' ./ /media/myusb/ I get the first image at url.
I have a tiny shell script to rsync files between two servers and remove the source files.
This script works fine, when it has been initiated manually or even when the rsync command is executed on the command line.
But the same script doesn't work, when I try to automate it through crontab.
I am using 'abc' user to execute this rsync, instead of root, as root login to servers are restricted in all of our servers, by us.
As I mentioned earlier, manual execution works like charm!
When this rsync.sh is initiated through crontab, it runs the first command(chown abc.abc ...) perfectly without any issues. But the second line is not at all executed, and there is no log entry i can find at /mnt/xyz/folder/rsync.log.
I just tried to sync files from one server to another. After the sync process, I found the files are bigger than original ones.
I looked up the web and found someone mentions the rsync daemon. So I have to run the daemon on one server before I run the rsync?
The command I used is rsync --partial --progress -r source destination
use rsync to cp such files and dirs under /var/www/html/mydir directory but these two files(/dir4/1.html /dir4/2.html) cant rsync to dest mechine.
rsync configure file,below...
I'm using Ubuntu 10.04 LTS server and Postgresql 8.4. I have a .sh script that is run by cron every other hour. That works fine. The .sh script includes an rsync command that copies a postgresql dump .tar file to a remote archive location via ssh. That fails when run by cron; I think because it is (quietly) asking for the remote user's password (and not getting it). I set up the public/private ssh key arrangement. The script succeeds when run manually as the same user that the cron job uses, and does not ask for the password. I am able to ssh to the remote server from the source server (using the same username) and not get the password prompt (both directions), so why doesn't rsync work? I even put a .pgpass file in the root of that user's directory with that user's password, and the user/password are identical on both servers.
I think the problem is rsync is not able to use the ssh key correctly. I tried adding this to my script but it didn't help.
Code:
Here is the rsync command embedding in the .sh script.
Code:
Here is the cron entry:
Code:
I just got a new server dedicated running php5 and SUExec. I understand the concept. INstead of running as "nobody", PHP scripts run as the user of that website. This also means I don't have to have some special directory where I set permissions to 775 or 770 with group=nobody etc just to upload files via PHP .. since PHP will upload files just as the user would using FTP. OK, great so far, and more security, I'm all for it.
View 1 Replies View RelatedNot sure what else to call it. My softlinks are not updating to reflect the parent/linked folder-files.
Code:
ln -s /folder/* /home/user/Music
This produces what I want. However, if I make any changes to /folder/, I dont see a reflected change in the links at /home/usr/Music.
A added plus would be to be able to use folder/file thumbnails from the source directory on the symlinks. Are sym links just shortcuts or are they real links to the directory (dynamic).
I have a directory full of symlinks. I want to zip them as if there were the real files. If I just zip them as usual, the resulting zip file does not contain the files at all! Is there any way to do this?
View 3 Replies View RelatedI'm wondering if there is a safe and reliable way of converting symlinks to directories to hardlinks? Basically I have a bunch of files in a directory tree under directory A, and I have in another directory B a set of symlinks to the original site. I would like all the files to end up in B and A to be empty. The links are all to directories, not to files, so the soft2hard script here won't work.
I don't have enough free diskspace to copy the files from A to B so I thought that converting the softlinks to hard links and then removing directory A would work? There are many directories, so would like to do it automatically if at all possible.
(In case anyone is intrigued, the issue is with my maildir directory. I used to use kmail, but with the move to kmail2 it became buggy and virtually unusable with large amounts of mail and I had to find another solution. I now use dovecot with gnus and mutt as MUA to access the maildirs. kmails maildirs are non-standard (!), so with the help of a script I found online, I created ~/dovecot and in it put symlinks to all the maldirs under ~/.kde/share/apps/kmail. The problem with this is if I accidentally start kmail it will try to access that directory and mess up all my tags, so I want to completely move all my mail from ~/.kde/share/apps/kmail to ~/dovecot, but without messing up all my tags etc.)
When I access my OpenSUSE box's samba shares from another PC I am able to browse them, but am not able to follow symlinks.I've had the problem before on Ubuntu and fixed it by adding the following lines to my /etc/samba/smb.conf
unix extensions = No
wide links = Yes
follow symlinks = Yes
I've added them into my OpenSUSE smb.conf (below) but still no joy.
I have the following share setup on my Ubuntu machine
# From /etc/fstab
//192.168.1.13/media /home/USER/SHARE/media cifs username=USERNAME,password=PASSWORD,_netdev,uid=USER,gid=users 0 0
[code]....
I recently tried the package "pysdm" to mount a data partition but didn't care for the way it handled things so I opted for editing fstab which I'm much more satisfied with.I did then purge pysdem and it's config files but something lingered, finally I got things working by using a UUID designation in fstab but I know there's an old symlink to a /dev/sdb7 that has since been deleted and later recreated with a new UUID.
I have "googled" enough to find that the package "fslint" should be able to handle this but since one gui got me in trouble I'm reluctant to use another, I'd prefer just being able to find where symlinks live and ripping them out by the short hairs
What is the syntax if you want to create a symlink to a resource on a mounted drive? am I missing something?
View 2 Replies View RelatedNot sure this is appropriate here but wondering if one has advantage over other in terms of sharability?
Also is it possible to upload files through symlinks? ie put a simlink to a tar ball located elsewhere on filesystem than ubuntu one/dropbox folder.
shed some light on what the behavior would be if I symlinked a non-encrypted directory (lets call it /media/sdcard/foo) into an encrypted home directory (lets say home/documents/sdcard-docs)Would files written to that symlinked directory be encryptedgarbled, or would they be normal? would files that were attempted to be read be interpreted as encrypted, or would they be read properly?
View 1 Replies View RelatedI did a "backup" of my .wine folder by copying it to my network attached storage drive. Now, that folder thinks it has about 200gb of data! Specifically, it's the .wine/dosdevices folder. I didn't grasp that Wine had created some kind of hard links to my ~ folder on my Ubuntu box. Since that folder is clearly recursively linking to all my files, I'm afraid to just delete it. I did delete one file within it as a test and, sure enough, the original file back on my Ubuntu box disappeared!how can I delete the .wine "backup" folder I created on my NAS without deleting all the original files?
View 5 Replies View RelatedI'm hosting some sites on a dedicated server. All these sites make use of the same codebase (couple of directories with php and javascript files). Because I want to manage this codebase on only one location, I was trying to share it via symlinks. When looking at commandline, everything seemed to work fine. However, when browsing to the sites that use the codebase, problems occur. It seems as if php cannot find the files.
View 3 Replies View RelatedI tried to run one grep that excludes symlinks and pipe it into a second that searches for my pattern but I continue to get the recursive directory loop error. As a result (I think) my grep is incomplete as the search just loops between directories symlinked together.
View 7 Replies View RelatedThis system was first installed on a 120Gb dual booted hard drive with Debian split with XP. It grew into a development system spanning 3 drives and 360Gb.
The addition drives were NOT mounted under the standard unix conventions, such as sdb being /usr and sdc being /home. They are mounted as /drive1, /drive2. All ext4 (except for the /dev/sda1 XP partition).
For example opt is symlinked to /drive1/opt
Firefox is at /drive1/opt/firefox.
So far so good, and with 20,000 packages installed, the system does work. With however, some caveats.
Certain programs like apt, refuse to recognize symlinks. I cannot move its /archives to the other drives. Some other stuff seem equally recalcitrant. Especially stuff from the lower init levels (that is understandable as they are running before the mounting process).
But others are unexpectedly recalcitrant.
Is there any way to *force* systemwide respect for symlinks?
Reducing size of XP is a non-starter here...
that's it! I got OpenSuse 11.2 and the size of the Repo has exceeded the 4 GB limit. This means that the mirror doesn't fit on my USB-stick anymore, nor would it fit on aDVD-RAM/RW (thus I can't distribute the updates anymore). Yes, there are bigger USB-sticks and BR-RWs, but in the end it is pure insanity!
View 1 Replies View RelatedLet's say you have this structure:
+ directory
-- file1
-- file2
-- file3 -> /tmp/file3
file3 is a link to another file3 somewhere else on the system.
Now let's say I chmod 777 the directory and all contents inside it. Does my file3 in /tmp receive those permissions? Also, let's say we have the same situation but reversed.
/tmp/file3 -> /directory/file3
If I apply the permissions on the file being linked to, how does that effect the link?
I just made the upgrade to 10.04 over the weekend, and everything seems to be working fine, minus one nagging detail.I have a Mythbuntu setup - a frontend and a backend. In addition to recording television, I have a folder setup on my backend where I can dump movie files to watch on the frontend. The folder is shared in the /etc/fstab via a CIFS share. Within that folder is a symlink to where my torrent folder is located. The issues seems to be how my frontend handles symlinks within the shared folders. If I run "ls -la" in my base CIFS share folder, it lists the symlink folder and says it's owned by root, but if I try and change to that directory - either using "ls" or "cd" - it says Permission Denied. On the backend, I changed ownership of the symlink to the username the CIFS is using to login, but that had no effect. I'm not familiar with how to do any sort of configuration on CIFS, if that's where I'd even need to start.
View 2 Replies View RelatedThis question is about windows 7. I want the "C:program Files" folder to be a symlink to a directory on another partition. I can't do this from within windows because it locked the Program Files Folder. Will a linux live CD do the trick?
View 1 Replies View RelatedI was about to post a new thread and saw that there are several answers out here already and I have done the new group permission one on my server -- actually on my test server I just added "my user" to the group "root" to gain rw access to /var/www/htdocs/.Someone suggested that the proper way might be to do symlinks to the directory in a real production environment and I wanted to find out if that is the *best* way to go or whether to actually make a group "www_admin" (pick your favorite flavor of this) and add my users to that group?
I guess I am looking for the "best practice" in a real corporate production environment that is most secure.
I wanted to use symlinks, in my annon ftp dir. I was running vsftp, but everyone says it can't be done. So I have changed to proftpd. I link to the annon ftp server on my website. I get the dir listing, with symlinks, but try to follow them I get error 550, it don't exist. I have read [URL], but after following the advice here. I get the same error. I have not tried the mount bind option, as I want symlinks or hard links either will do.
View 3 Replies View Relatedthe hard drive on my web server died so I replaced it and installed Fedora 15. I have been installing the OS on a small drive and then mounting other drive(s) for storage and did the same thing here. The problem is that I cannot get Apache to follow symlinks. Apache is partially working, it will serve my index.html page in the web page root directory and I made another directory called webtest and put some files in, and that worked fine. I made a symlink from the root directory to a folder on the mounted drive but when I try to access it, it gives me the old 403 Forbidden error. I also made a similar link in the webtest directory, I can view the other files in that directory but it does not show the link - also, trying to force it by typing the full path in the browser gives the same 403 error.
I have checked the permissions and SELinux types (using ls -Z) and they seem to be fine and I cannot find any .htaccess files. All the files on the mounted drive are owned by www:www but I also tried setting the user and group to www in httpd.conf and it still gave the same issue, so I changed them back to apache. I've been reading posts and sites on this for hours now and I'm still lost. I can post results of log files if they are of any use, I just wasn't sure what to post because it is somewhat working.
I have a samba mount in Windows, mounted from a local Linux devbox. The projects on this mount contain a lot of symlinked directories. They all work fine and look like normal directories to Windows. What I'd like to do is figure out for a given file what its real UNIX path is.The context is: in my editor I have a script that runs the Perforce open command, to which I pass the file location.p4.exe edit FILE_LOCATION
However, perforce will not understand this FILE_LOCATION, if on the Linux side part of the path is a symlink. It needs to know the real path.What I need is something along the lines of the Linux readlink command, which resolves symlinks in paths.
I'm trying to get my music folder into something sensible. Right now, I have all my music stored in /home/foo so I have all of the albums soft linked to ~/music. I want the structure to be ~/music/<artist>/<album> I've got all of the symlinks into ~/music right now so I just need to get the symlinks into the proper structure. I'm trying to do this by delving into the symlinked album, getting the artist name with id3info. I can do this, but I can't seem to get it to work correctly.
for i in $( find -L $i -name "*.mp3" -printf "%h
")
do
echo "$i" #testing purposes
[code]....
I'm trying to mv The Trance Mix 3 Wanderlust - DJ Exortius [TRANCE DEEP VOCAL TECH]@ into the real directory DJ Exortius. DJ Exortius already exists, so it's just a matter of moving it into the correct directory that's based on the id3 tag of the mp3 inside.
PS: I've tried easytag, but when I restructure the album, it moves it from /home/foo which is not what I want.