dump/restore of a fs with 100 of millions hardlinks


 
Thread Tools Search this Thread
Top Forums UNIX for Dummies Questions & Answers dump/restore of a fs with 100 of millions hardlinks
# 1  
Old 11-12-2009
dump/restore of a fs with 100 of millions hardlinks

Hi :-)

i have a dump of a backupdisk (~540GB / ext3). The Backups have some 100 millions of hardlinks (backups are created with storeBackup). The OS is linux.

A restore of a directory ended after some days with the errormessage "no memory to extend symbol table"

The restore of the complete filesystem ended with a restore af 8GB data and the message Out of Memory: Kill process 2618 (restore) score 73 and children. Out of memory: Killed process 2618 (restore)."

The testsystem has a debian minimal installation and 650MB Ram. Of course I can try it on an other system and wait some days until I have maybe the same message.

An interactive Restore crashed after some days of waiting :-(

codeexample how I restore the whole FS
# restore rd -f filename.dump
in the /tmp dir, the rstdirfile is 5.7 GB and the rstmode file 560M. Here I have the out of memory problem

codeexample how I restore one directory (with all the files)
# restore -x daten -f filename.dmp
here I have the "no memory to extend symbol table" problem ...

First question: is dump/restore aible to handle this numbers of files?
2nd question: if yes, what I make wrong?

greeting from berlin
turricum
 
Login or Register to Ask a Question

Previous Thread | Next Thread

10 More Discussions You Might Find Interesting

1. Solaris

restore dump from veritas format using ufs file system

hi all i have a DLT tape in that tape backup is there is in veritas volume format and i want to restore it in ufs file system how can i do it? right now i don't have veritas file system setup. i have only ufs file sysytem please help some production data is to be restore. (3 Replies)
Discussion started by: nikhil kasar
3 Replies

2. Shell Programming and Scripting

HardLinks and Softlinks

How do i make a hardlink readable,writable, and executable by me? I am kinda hoping for a command i can use Also, i was wondering what file type are hardlinks and softlinks?Are they directories? (10 Replies)
Discussion started by: lilbo4231
10 Replies

3. UNIX for Dummies Questions & Answers

Dump and Restore

I want to copy over the contents of one partition on hard drive A to another partition on hard drive B. Using DD command caused problems since it overwrote the partition information on hard drive B and made hard drive B's partition size the same as the one on partition A. Our department manager... (5 Replies)
Discussion started by: mojoman
5 Replies

4. UNIX for Dummies Questions & Answers

How to restore a dump file on the disk

Hi all, i am a real dummy to unix and in need of help.My platform is Sun solaris(5.9) I have a dump file, an oracle cold backup taken with ufsdump command. This dump file resides on the disk, not the tape. I want to extract this dump file to a directory. But i cant, i read about ufsrestore... (1 Reply)
Discussion started by: merope
1 Replies

5. UNIX for Dummies Questions & Answers

Dump and restore problem

Hi all, having problem with dump & restore for cloning simple linux system. The command is : dump -0u -f /temp/var.dump /dev/sda5 But somehow the result displayed: ~~~~~~ dump: ACLs in inode #1250760 won't be dumped: Inappropriate ioctl for device dump: ACLs in inode #1250761 won't be... (1 Reply)
Discussion started by: mycoolwater
1 Replies

6. Red Hat

Dump and restore command usage ??

Hi, I am using RHEL 4.0 I need to take backup of a directory and then restore it to some other location. For taking Backup of final directory, I am using this code: dump -0aj -f /home/vicky/final.dump /home/vicky/final/ Now, I am trying to restore this final.dump to some other... (2 Replies)
Discussion started by: vikas027
2 Replies

7. UNIX for Dummies Questions & Answers

How to list hardlinks?

Q1: Let's say I create a hard-link bar.c in /tmp to a file foo.c which resides in /var/tmp. Is there a easy way to find out which file /tmp/bar.c hardlinks to (and vice-versa - i.e which files have got hard-linked from /var/tmp/foo.c) when one does not (and wants to) know the location of the other... (0 Replies)
Discussion started by: mahatma
0 Replies

8. Shell Programming and Scripting

restore mysql dump file in many remote servers?

Hi all, I want to restore DB file in many mysql servers, i already using script for sending the dumpfile in all servers, but it's just annoying if i have to restore the dumpfile in all servers, i want just execute 1 script, and will restore in all remote mysql servers. I make script but not... (2 Replies)
Discussion started by: blesets
2 Replies

9. UNIX for Dummies Questions & Answers

restore dump - help

I have a file with .dmp.gz. I am trying to uncompress this file in a new directory. was able to use gzip to take off the gz extension. How do I get the .dmp extension off? Will ufsrestore work on SUN system? HELP! (2 Replies)
Discussion started by: sankera
2 Replies

10. UNIX for Dummies Questions & Answers

search for hardlinks based on filename via find command

I am using command substitution into a find command in a script where I have built a menu to do a bunch of tasks within my unix account. When I choose the options for to find a file/files that have the same inode of the entered filename, ie hardlinks, nothing shows up. When I choose the appropiate... (2 Replies)
Discussion started by: hunternjb
2 Replies
Login or Register to Ask a Question
EXT3GREP(8)						 recover files on ext3 filesystems					       EXT3GREP(8)

NAME
ext3grep - ext3 file recovery tool SYNOPSIS
ext3grep [OPTIONS] FILE ... DESCRIPTION
ext3grep is a simple tool intended to aid anyone who accidentally deletes a file on an ext3 filesystem, only to find that they wanted it shortly thereafter. OPTIONS
--superblock Print contents of superblock in addition to the rest. If no action is specified then this option is implied. --print Print contents of block or inode, if any. --ls Print directories, one line per entry. See the FILTERS section for details on filtering this output. --accept FILE Accepts 'file' as a legal filename. Can be used multiple times. If you change any --accept you must remove BOTH stage* files! --journal Prints the contents of the journal. --show-path-inodes Show the inode of each directory component in paths. Filters: --group gid Only show/process files owned by process group gid. --directory Only show/process process directory inodes. --after dtime Only show/process entries deleted on or after dtime. --before dtime Only show/process entries deleted before dtime. --deleted Only show/process deleted entries. --allocated Only show/process allocated inodes/blocks. --unallocated Only show/process unallocated inodes/blocks. --reallocated Do not suppress entries with reallocated inodes. Inodes are considered 'reallocated' if the entry is deleted but the inode is allo- cated, but also when the file type in the dir entry and the inode are different. --zeroed-inodes Do not suppress entries with zeroed inodes. Linked entries are always shown, regardless of this option. --depth depth Process directories recursively up till a depth of 'depth'. Actions: --inode-to-block inode_num Print the block that contains inode inode_num. --inode inode_num Show info on inode inode_num. If --ls is used and the inode is a directory, then the filters apply to the entries of the directory. If you do not use --ls then --print is implied. --block block_num Show info on block block_num. If --ls is used and the block is the first block of a directory, then the filters apply to entries of the directory. If you do not use --ls then --print is implied. --histogram=[atime|ctime|mtime|dtime|group] Generate a histogram based on the given specs. Using atime, ctime or mtime will change the meaning of --after and --before to those times. --journal-block block_num Show info on journal block block_num. --journal-transaction seq Show info on transaction with sequence number seq. --dump-names Write the paths of files to stdout. This implies --ls but suppresses its output. --search-start str Find blocks that start with the fixed string str. --search str Find blocks that contain the fixed string str. --search-inode block_num Find inodes that refer to block block_num. --search-zeroed-inodes Return allocated inode table entries that are zeroed. --inode-dirblock-table dir Print a table for directory path dir of directory block numbers found and the inodes used for each file. --show-journal-inodes inode_num Show copies of inode inode_num still in the journal. --restore-file path Will restore file path. path is relative to root of the partition and does not start with a '/' (it must be one of the paths returned by --dump-names). The restored directory, file or symbolic link is created in the current directory as ./path. --restore-all As --restore-file but attempts to restore everything. The use of --after is highly recommended because the attempt to restore very old files will only result in them being hard linked to a more recently deleted file and as such pollute the output. --show-hardlinks Show all inodes that are shared by two or more files. --version, -[vV] Prints the version information and exits. --help, Prints a help message and exits. EXAMPLES
Restoring all files from the ext3 partition/file /backup/sda1: ext3grep --restore-all /backup/sda1 Listing the files owned by GID 1000 on /backup/sda1: ext3grep --ls --group 1000 /backup/sda1 Finding all files containing the string Critical_report in their name on /backup/sda1: ext3grep --dump-names /backup/sda1 | grep 'Critical_report' LIMITATIONS
Do not attempt to use ext3grep for recovery from a mounted filesystem. Ever. No, not even then. ext3grep sometimes runs out of memory spare on 32-bit architectures and crashes. It is highly recommended that you run ext3grep in a 64-bit environment when dealing with large filesystems, though this is seen as a bug. ext3grep cannot recover files if there are no remnants of them. Some files that ext3grep recovers may have trailing null bytes - just scrape them off like the burnt bits on toast. SEE ALSO
debugfs(8) AUTHOR
ext3grep was written by Carlo Wood <carlo@alinoe.com>. This manual page was written by Rich Ercolani <rercola@acm.jhu.edu>, for the Debian project (but may be used by others). It may be distrib- uted under the same terms as ext3grep, the GNU General Public License, either version 2 or (at your option) any later version. 0.8.0 2008-09-14 EXT3GREP(8)