hardlink - Page 2 - Developer IT

How can I find all hardlinked files on a filesystem?

- by haimg

I need to find all hardlinked files on a given filesystem. E.g. get a list of files, each line contains linked pairs, or triplets, etc. I understand more or less how to do it, one needs to create a dictionary keyed by inode for all files/directories on a filesystem, exclude "." and ".." links, and then indodes with more than one name are hardlinks... But I hope that maybe a ready-made solution exists, or someone already wrote such a script.

Read the article

How to properly store dotfiles in a centralized git repository

- by asmeurer

I'd like to put all my dotfiles (like .profile, .gitconfig, etc.) in a central git repository, so I can more easily keep track of the changes. I did this, but I would like to know how to properly handle keeping them in sync with the actual ones in ~/. I thought that you could just hard link the two using ln, but this does not seem to work as I expected, i.e., if I edit one file, the other does not change. Maybe I misused the ln command, or else I misunderstand how hard links work. How do people usually do this? Judging by GitHub, it's a pretty popular thing to do, so surely there's a seamless way to do it that someone has come up with. By the way, I'm on Mac OS X 10.6.

Read the article

Break a hard link of a file in use

- by Stebi

I used hard links to merge duplicated files on my SSD (space is still precious) and now have a weird problem. Common files like msvcr110.dll got hard linked. Now I want to delete a program which has this file in its installation directory. But I cannot because this file (on another location) is used by a currently running application (don't know which) and windows doesn't allow me to delete this file because it's in use. I can rename the file but it still points to the same file, so not possible to delete it. Is there any way to break a the hard link of a file which is currently in use? I currently use a trash folder where I move those files to so I can delete the directory structure of program to be deleted. But I'd like to get rid of this leftover (although it doesn't take much space as it's a hard link).

Read the article

How much space do NTFS hardlinks/symlinks occupy?

- by Felix Dombek

Well, I guess it must be something proportional to the original filename plus the new filename for symlinks, and only the new filename for hardlinks, but how does this affect the disk space exactly? I just made a folder with about a hundred thousand symbolic links in it, and the folder still reported 0 bytes usage. I may be mistaken, but I even think the free capacity of the drive remained the same. Then I permanently deleted the folder and the sizes still stayed the same. Could I fill up a hard disk just with symlinks? Or does NTFS have limitations in that no more than x symlinks are allowed on one drive/in one folder, so the capacity of the drive cannot be reached?

Read the article

All of the NTFS hard links damaged, where are hardlinks stored and how to recover them?

- by String Xu

This is Windows 7 x64 sp1 on a NTFS file system. All hardlinks within C:\Windows\System32 folder disappear, and the Windows can't boot, because even the osloader, C:\Windows\System32\boot\Winload.exe also disappeared. Nevertheless, the original files are still located in the corresponding C:\Windows\winsxs folders. After booting into the Recovery Environment, and copied one Winload.exe (x64) from other folder, Windows gave an error pointing out that "ntoskrnl.exe is corrupted or missing...its file digital signature cannot be verified" In trying to boot in Safe Mode, the message above was shown after a screen prompting "Loaded \Windows\system32\config\system" Because at this early booting stage, smss.exe was still not loaded, so there is not any dumping and logs. Based on my study, ntoskrnl.exe depends on the following files: C:\windows\system32\PSHED.DLL C:\Windows\System32\hal.dll C:\Windows\System32\kdcom.dll C:\Windows\System32\clfs.sys C:\Windows\System32\ci.dll All those files above are copied from their corresponding folders and verified their md5 with a well-operating Windows 7 x64 SP1. But the booting error is still the same: "ntoskrnl.exe is corrupted or missing..." Background: 1. Before the reboot, there was an windows update going on. Then something unknown happen, almost all processes were broken to run, including the windows task manager, taskmgr.exe. After mount the hard disk to other computer, it seems that all hardlinks within C:\Windows\System32 folder were gone. I tried several data recovery software, but they are not be able to find those disappeared NTFS hard links. So the question is: Where are information about those hard links stored? And how to recover them? Are they depend on some windows service or stored in the registry?

Read the article

All of the NTFS hard links disappear, where are hardlinks stored on disk and how to recover them?

- by Osiris

This is Windows 7 x64 sp1 on a NTFS file system. All hardlinks within C:\Windows\System32 folder disappear, and the Windows can't boot, because even the osloader, C:\Windows\System32\boot\Winload.exe also disappeared. Nevertheless, the original files are still located in the corresponding C:\Windows\winsxs folders. After booting into the Recovery Environment, and copied one Winload.exe (x64) from other folder, Windows gave an error pointing out that "ntoskrnl.exe is corrupted or missing...its file digital signature cannot be verified" In trying to boot in Safe Mode, the message above was shown after a screen prompting "Loaded \Windows\system32\config\system" Because at this early booting stage, smss.exe was still not loaded, so there is not any dumping and logs. Based on my study, ntoskrnl.exe depends on the following files: C:\\windows\\system32\\PSHED.DLL C:\\Windows\\System32\\hal.dll C:\\Windows\\System32\\kdcom.dll C:\\Windows\\System32\\clfs.sys C:\\Windows\\System32\\ci.dll All those files above are copied from their corresponding folders and verified their md5 with a well-operating Windows 7 x64 SP1. But the booting error is still the same: "ntoskrnl.exe is corrupted or missing..." **Background:** Before the reboot, there was an windows update going on. Then something unknown happen, almost all processes were broken to run, including the windows task manager, taskmgr.exe. After mount the hard disk to other computer, it seems that all hardlinks within C:\Windows\System32 folder were gone. I tried several data recovery software, but they are not be able to find those disappeared NTFS hard links. So the question is: Where are information about those hard links stored? And how to recover them? Are they depend on some windows service or stored in the registry?

Read the article

Synchronize folder on network, preserving hard links

- by Waleed Hamra

I have few computers using Windows XP Pro. I want to synchronize/back a folder from one machine, to another one. This far, It's a simple problem, and I've used FreeFileSync for such operations, with very satisfactory results. But, this all changes when hard links come into play. Today's folder contains lots of hard links, using such backup programs will result in hard links being treated as multiple files, and copied as such, greatly increasing folder size on destination, and defeating the purpose of using all these hard links in the first place. It gets more complicated when we take into consideration the fact that network shares on Windows DON'T expose hard linking facilities, meaning that running a hard-link-aware tool like rsync using --hard-links will be of no use. So my question, how can i backup my folder to the other computer, while preserving hard links? I don't mind installing 3rd party tools to do it, as obviously, the standard windows shares approach won't work... I am guessing there might be some tool that can be installed on both machines and works in a server/client mode? anyone has any idea how to do this?

Read the article

Backing up an rsnapshot directory to a remote device

- by user123480

I have a local backup server that uses rsnapshot with hard links that contains about 10TBs of information which add about 4 to 5GBs per day. It's works great. I've been requested to set up and maintain a remote backup of the local rsnapshot directory structure. It's a nightly backup. I've tried using rsync with encryption which takes forever and eats system resources. A previous post says not to use rsync with hard links for that reason. I need a suggestion of how I can keep the local and remote copies of the rsnapshot structures in sync? Thanks

Read the article

How can I check the actual size used in an NTFS directory with many hardlinks?

- by kbyrd

On a Win7 NTFS volume, I'm using cwrsync which supports --link-dest correctly to create "snapshot" type backups. So I have: z:\backups\2010-11-28\cygdrive\c\Users\... z:\backups\2010-12-02\cygdrive\c\Users\... The content of 2010-12-02 is mostly hardlinks back to files in the 2010-11-28 directory, but there are a few new or changed files only in 2010-12-02. On linux, the 'du' utility will tell me the actual size taken by each incremental snapshot. On Windows, explorer and du under cygwin are both fooled by hardlinks and shows 2010-12-02 taking up a little more space than 2010-11-28. Is there a Windows utility that will show the correct space acutally used?

Read the article

Efficient mirroring of directories using hardlinks

- by zoqaeski

I'm backing up my music collection on to a number of NTFS-formatted external hard-drives; however, as I store my main collection in FLAC and have my library on my laptop as MP3s to save space, I want to be able to back up both sets, because mass conversion between formats is time-consuming. The "music" directory can contain any format; the "mp3s" directory contains only MP3s converted from files in the "music" directory. The music collection on the laptop contains only MP3s, but they come from both sources. When I backup my laptop's library to the "mp3s" directory, I want to only copy across MP3 files that don't exist in the "music" directory; those that do should be hard-linked to the "music" directory. All directories have an identical hierarchy, sorted by artist, album, date, discnumber if applicable, etc, and I use a tagging editor to ensure consistency across all these locations. I'm also using a Linux computer, but keeping the music collections on NTFS-formatted partitions so that they are readable by both Linux and Windows. At the moment, I use the following command to perform the backups, but this is time-consuming due to the expensive nature of finding hard links. rsync -avu --progress --relative --ignore-existing --link-dest=../music/ **/*.mp3 /media/ntfspocket/mp3s Is there a way to perform this backup more efficiently, taking advantage of the directory hierarchy?

Read the article

Is there a way to create a copy-on-write copy of a directory?

- by BCS

I'm thinking of a situation where I would have something that creates a copy of a directory, tweaks a few files, and then does some processing on the result. This wold be done fairly often, maybe a few dozen times a day. (The exact use case is testing patch submissions; dupe the code, patch it, build/test/report/etc.) What I'm looking for could be done by creating a new directory structure and populating it with hard links from the origonal. However this only works if all the tools you use delete and recreate files rather than edit them in place. Is there a way to have the file system do copy-on-write for a file? Note: I'm aware that many FSs use COW at a block level (all updates are done via writes to new blocks) but this is not what I want.

Read the article

hardlinks on ntfs with windows

- by knittl

is it possible to create hardlinks for a file on an ntfs partition using windows? ntfs obviously can handle hardlinks, since creating them with ntfs-3g works – the links even work in windows. or is this the only way to create hardlinks on ntfs?

Read the article

Efficient mirroring of directories using hard links [closed]

- by zoqaeski

I'm backing up my music collection on to a number of NTFS-formatted external hard-drives; however, as I store my main collection in FLAC and have my library on my laptop as MP3s to save space, I want to be able to back up both sets, because mass conversion between formats is time-consuming. The "music" directory can contain any format; the "mp3s" directory contains only MP3s converted from files in the "music" directory. The music collection on the laptop contains only MP3s, but they come from both sources. When I backup my laptop's library to the "mp3s" directory, I want to only copy across MP3 files that don't exist in the "music" directory; those that do should be hard-linked to the "music" directory. All directories have an identical hierarchy, sorted by artist, album, date, discnumber if applicable, etc, and I use a tagging editor to ensure consistency across all these locations. I'm also using a Linux computer, but keeping the music collections on NTFS-formatted partitions so that they are readable by both Linux and Windows. At the moment, I use the following command to perform the backups, but this is time-consuming due to the expensive nature of finding hard links. rsync -avu --progress --relative --ignore-existing --link-dest=../music/ **/*.mp3 /media/ntfspocket/mp3s Is there a way to perform this backup more efficiently, taking advantage of the directory hierarchy?

Read the article

Hard link not works under MacOS in GUI mode

- by AntonAL

Hi, i faced a little strange behavior, while using hard links. From terminal, i create a text file 1.txt and a hard link "to this file" nano 1.txt mkdir dir ln 1.txt ./dir/ I check the resulting hard link and see, that it's contents is the same, as of "original" file. less ./dir/1.txt I change the initial file ... nano 1.txt ... and see, that changes was reflected in hard-link less ./dir/1.txt I change content of hard-link (more correct, of course - file, being referenced with hard-link) ... nano ./dir/1.txt ... and see, that changes are reflected in initial file less 1.txt Until now, all going well... Now, I close terminal and start playing with created files (1.txt and ./dir/1.txt) from Finder. When i change on this two files with TextEdit, changes are not reflected in another file. Just like the hard link was teared off... Whats going on here ?

Read the article

Delete link to file without clearing readonly bit

- by Joshua

I have a set of files with multiple links to them. The files are owned by TFS source control but other links to them are made to them. How do I delete the additional links without clearing the readonly bit. It's safe to assume: The files have more than one link to them You are not deleting the name owned by TFS There are no potential race conditions You have ACL full control for the files The machine will not lose power, nor will your program be killed unless it takes way too long. It's not safe to assume: The readonly bit is set (don't set it if its not) You can leave the readonly bit clear if you encounter an error and it was initially set Do not migrate to superuser -- if migrated there the answer is impossible because no standard tool can do this.

Read the article

Creating directory hard links in MacOS X?

- by felixge

How can I create a hard link to a directory in OSX? This feature has been added to their file system in 10.5 (for time machine), but I could not find any information on actually using it from the command line.

Read the article

css issue on hover- shaky effect

- by Sarika Thapaliya

Read the article

Backup to Synology NAS using rsync or NFS and hardlinks

- by danilo

I want to backup data from a Windows (Vista) computer to a Synology NAS (210j). The NAS supports FTP, SMB, NFS and also allows a rsync daemon to be set up. I want to backup different folders to the NAS, but I'd prefer to use the hardlink method to save diskspace (like this script does). With this method, a new folder is created for every backup, but if the file already exists on the target, only a hardlink is created. The filesystem on the Synology device is ext3, so I probably can't use rsyncbackup, as it is made for NTFS. Is there another way to do this backup with hardlink support?

Read the article

Unable to start backuppc

- by iUngi

I had a ext4 drive today I replaced with a RAID drive, I moved all the files from the old HDD. After that I tried to start the Backuppc but I'm getting the following error: Can't create a test hardlink between a file in /media/WESYS_RAID/backups/nbackuppc/pc and /media/WESYS_RAID/backups/nbackuppc/cpool. Either these are different file systems, or this file system doesn't support hardlinks, or these directories don't exist, or there is a permissions problem, or the file system is out of inodes or full. Use df, df -i, and ls -ld to check each of these possibilities. Quitting... The permission looks like this: drwxrwxrwx 2 backuppc backuppc 4096 2012-04-12 11:06 cpool -rwxrwxrwx 1 backuppc backuppc 14290 2012-04-10 16:00 dead.letter drwxrwxrwx 2 backuppc backuppc 4096 2012-04-12 10:56 log drwxrwxrwx 2 backuppc backuppc 4096 2012-04-12 11:08 pc drwxrwxrwx 2 backuppc backuppc 4096 2011-10-27 22:40 pool drwxrwxrwx 2 backuppc backuppc 4096 2011-10-27 22:40 trash I also tried to create hardlinks and it does work ln -i test.txt testlink The result of the df -i Filesystem Inodes IUsed IFree IUse% Mounted on /dev/sdc 122101760 1279684 120822076 2% /media/WESYS_RAID /dev/mapper/WeSyS_LVM 115687424 308565 115378859 1% /media/WESYS_LVM What am I doing wrong?

Read the article

Webserver: chrooted PHP gives mysql.sock error when attempting to reach mysql

- by Jon L.

Hey guys, I've configured an Ubuntu webserver with Nginx + PHP5-FPM. I've created a chrooted environment (using jailkit) that I'm tossing my developers into, from where they can develop their test applications. Chroot jail: /home/jail Nginx and PHP5-FPM run outside the chroot, but are configured to function with websites within the chrooted environment. So far, Nginx and PHP5-FPM are serving up files without issue, except for the following: When attempting to connect to MySQL, we receive this error: SQLSTATE[HY000] [2002] Can't connect to local MySQL server through socket '/var/run/mysqld/mysqld.sock' Now, I believe the issue is due to the non-chrooted php.ini referencing mysqld.sock outside of the chroot environment (it's actually using the MySQL default setting currently). My question is, how can I configure PHP to access MySQL via loopback or similar? (Found that as a suggestion in a google result, but without any instructions) Or if I'm missing some other obvious setting, let me know. If there's an option of creating a hardlink (that would remain available even if mysql is restarted), that would be handy as well.

Read the article

ubuntu one syncronizing problems

- by user72249

I am a user of Ubuntu and Ubuntu One. First: I had a serious problem with ubuntu one. I have a windows and several Ubuntu machine and they all sync one account. My first problem is that i can't really delete files. Whenever I delete something Ubu One resync it, so it appears again in my folder. Furthermore, when i move something to another dir it synconize the new location and resync the old one, so i got doubled my files. I can't reorganize my files. So i tried to delete the duplicated files through the web dash, but i cant reorder and select multiple files by extension. For example i wanted to move all PDF to another location... Second: I can't configure in the ubuntu one app the loacation of the main One folder. For example i want my One folder to be on another partition than my HOME folder. It is a main problem in linux and windows also. I tried to move that folder in windows with the hardlink method. So i uninstalled the U1, then i created a folder link to another drive, than installed the U1. THEN i lost all my files in my U1. Lucky thing that i have a backup, but i thought U1 is a stable solution for me, and i planned to extend my space, but these bugs are major problems! It's worth to pay for it if it's working perfectly. I think it's more important than releasing a new version in every month.

Read the article

Packaging MATLAB (or, more generally, a large binary, proprietary piece of software)

- by nfirvine

I'm trying to package MATLAB for internal distribution, but this could apply to any piece of software with the same architecture. In fact, I'm packaging multiple releases of MATLAB to be installed concurrently. Key things Very large installation size (~4 GB) Composed of a core, and several plugins (toolboxes) Initially, I created a single "source" package (matlab2011b) that builds several .debs (mainly matlab2011b-core and matlab2011b-toolbox-* for each toolbox). The control file is just the standard all: dh $@ There is no Makefile; only copying files. I use a number of debian/*.install files to specify files to copy from a copy of an installation to /usr/lib/. The problem is, every time I build the thing (say, to make a correction to the core package), it recopies every file listed in the *.install file to e.g debian/$packagename/usr/ (the build phase), and then has to bundle that into a .deb file. It takes a long time, on the order of hours, and is doing a lot of extra work. So my questions are: Can you make dh_install do a hardlink copy (like cp -l) to save time? (AFAICT from the man page, no.) Maybe I should just get it to do this in the Makefile? (That's gonna b e big Makefile.) Can you make debuild only rebuild .debs that need rebuilding? Or specify which .debs to rebuild? Is my approach completely stupid? Should I break each of the toolboxes into its own source package too? (I'll have to do some silly templating or something, because there's hundreds of them. :/)

Read the article

rsync not copying hard links

- by A.Ellett

I have two computers (both MacBook Airs) for which I sync one directory tree in both, but not the entire hard drive or any other directories. Let's say on computer A the directory is /Users/aellett/projects Let's say on computer B the directory is /Users/bellett/projects Generally, I'll log into computer B and then remotely connect to computer A as user 'aellett'. As super user I sync the two project directories as follows: rsync -av /Volumes/aellett/projects/ /Users/bellett/projects/ and this works as expected. On both computers I have another file letter.txt in a different directory which is not getting synced. Let's say on computer A the file is found in /Users/aellett/letters On computer B the file is found in /Users/bellett/correspondence Generally, I don't want to share what's not included in /Users/<username>/projects. But I do want to share this particular file. So on both computer I made a correspondence directory in projects. And then I made hard links as follows On computer A: ln /Users/aellett/letters/letter.txt /Users/aellett/projects/correspondence/letter.txt On computer B: ln /Users/bellett/correspondence/letter.txt /Users/aellett/projects/correspondence/letter.txt The next time I synced the two computers I did the following rsync -av -H /Volumes/aellett/projects/ /Users/bellett/projects/ When I checked on computer B, /Users/bellett/projects/correspondence/letter.txt was correctly synced. But, the hardlink to /Users/bellett/correspondence/letter.txt was no longer there. In other words, /Users/bellett/projects/correspondence/letter.txt was identical to /Users/aellett/projects/correspondence/letter.txt but it differed from /Users/bellett/correspondence/letter.txt. Since these two files were hard linked on both computers, I expected them to still have the hard link. Why are my hard links not being preserved?

Read the article

Copy past speed very slow for a large number of tiny files on Windows but not on linux

- by Arno2501

I've got this folder which contains 15'000 of tiny images (around 400 bytes each). If I copy past this folder on my laptop (Windows 7, i7 latest gen, superfast ssd) it takes about 30 seconds (yes for 7 megs !!!) the average transfer rate is 400 KBytes / second which is so slow. I mean my usual transfer rate is more like hundreds of MBytes per second !!! I get the same problem on my servers (Windows 2003, 2008 /r2) and on every Windows box that I could get my hands on. On the other hand if I do the same on a linux box (debian base, Ext3 FS) (which runs on the same SAN than all the windows servers I've tested) It's nearly instantaneous !!! I'm pretty sure the size / number of the files may stress such filesystem more than another but such differences !? Why is that ? Why is it so slow on the windows boxes (more that 30 sec for 7 MB) and so fast on the linux ones (one sec or so) (I mean this was not a hardlink that I've created it was a true copy). Is it a normal behaviour or something unusual ?

Read the article

Reading log files from web application

- by Egorinsk

I want to write a small PHP application for monitoring logs on a Debian server, including syslog logs and Apache/PHP messages. The problem here is that Apache user (www-data) has no access to /var/log directory. What would be the best way to grant an access to logs for PHP application? Let's assume that log files can be really large, like hundreds of megabytes. I have some ideas: Write a shell script that would be run via sudo and tail last 512 Kb of log into a separate file that can be read by application - that's ineffective, because of forking a new process and having to read data twice Add www-data to adm group (that can read logs) - that's insecure Start a PHP process via cron every minute to read logs — that's not very good, because it doesn't allow real-time monitoring. Also, this script will be started even when I don't read logs, and consume CPU time (server is in the cloud, and I'll have to pay for it) Create a hardlink for all log files with lowered permissions - I guess, that won't work because logrotate could recreate log files and they'll change inode number. Start a separate nginx/Apache server under privileged user that may read logs. Maybe anyone got a better solution?

Search Results

Search found 52 results on 3 pages for 'hardlink'.

Page 2/3 | < Previous Page | 1 2 3 | Next Page >

- by haimg

- by asmeurer

- by Stebi

- by Felix Dombek

- by String Xu

- by Osiris

- by Waleed Hamra

- by user123480

- by kbyrd

- by zoqaeski

- by BCS

- by knittl

- by zoqaeski

- by AntonAL

- by Joshua

- by felixge

- by Sarika Thapaliya

- by danilo

- by iUngi

- by Jon L.

- by user72249

- by nfirvine

- by A.Ellett

- by Arno2501

- by Egorinsk

< Previous Page | 1 2 3 | Next Page >