|
From: | rdiff |
Subject: | Re: [rdiff-backup-users] CRC error |
Date: | Wed, 7 Nov 2007 21:33:13 -0800 (PST) |
I'm not sure where this original post went to in the archives, but we have had the same problem. Below are 2 reposted messages on the subject from my sent-mail which discuss resolution. Hopefully they will find their way to the archive or perhaps be easier to find with your message below this one.
-Eric
From address@hidden Mon Jan 22 13:35:34 2007 -0800
Date: Mon, 22 Jan 2007 13:35:34 -0800 (PST) From: address@hidden To: address@hidden Subject: CRC check Failed Message-ID: <address@hidden> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed Status: O X-Status: X-Keywords: X-UID: 4793 We are using 1.0.4 and it has been running stable for many months. Every once in a while this CRC check comes across and in the past, the only way we knew to resolve it is to remove the rdiff-backup-data directory and start over. --check-destination-dir also fails with the same error. After reading the mailing list, they suggest that one of the .gz files is bad. After something like this: (http://www.mail-archive.com/address@hidden/msg01116.html) find dest/rdiff-backup-data -type f -name \*.gz -print0 | \ xargs -0r gzip --test I found that the file dest/rdiff-backup-data/mirror_metadata.2007-01-19T18:30:23-08:00.snapshot.gz was bad. The previous mirror_metadata file "mirror_metadata.2007-01-19T05:52:24-08:00.snapshot.gz" was intact so I copied the good one over the bad one and re-ran --check-destination-dir. Even though I used an older mirror_metadata, --check-destination-dir returned without error. I realize that because I copied an old metadata over a more recent backup's metadata some of the increments may be clobbered or perhaps confused. However, it is possible that the damaged metadata was a backup which never really did anything in which case, no harm done; unfortunately, this is unknown.What kind of metadata confusion/increment errors/restore errors might come up due to this metadata replacement which we should be aware of?
Is there a way to check the integrity of the increments? Can rdiff-backup perform restore operations on each file with increments and compare against some metadata md5/sha? Is there a way to keep mirror_metadata's from getting damaged or making the code which handles mirror_metadata more robust to failures? Does anyone have suggestions to avoid these scenarios in the future? Thank you for your help!!! -Eric ================================================ From: address@hidden To: Guenevere Prawiroatmodjo <address@hidden> cc: address@hidden, address@hidden Subject: Re: [rdiff-backup-users] Exception 'CRC check failed' In-Reply-To: <address@hidden> Message-ID: <address@hidden> References: <address@hidden> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed Status: O X-Status: X-Keywords: X-UID: 4813 On Fri, 26 Jan 2007, Guenevere Prawiroatmodjo wrote:
Hello,My last backup failed because of a hardware error, which probably caused
some
corrupt files in the process. When I use --check-desination-dir on the
backup
folder the following error msg appears:
This problem has come up many times on the mailing list, and so far there is not a good fix for it. DEVELOPERS: If you can think of a good fix which could be added to --check-destination-dir, this would be wonderful! I don't know enough about the inner workings of the rdiff-backup-data tree, but usually this error appears when one of the .gz files was incomplete. Try this: find rdiff-backup-data/ -name '*.gz" | while read a; do echo -n "$a:" gzip --test "$a" done This will tell you which files are bad. In our case, we had a problem with the mirror_metadata files. Search the archive for specifics, but basically I copied a good mirror_metadata over the bad mirror_metadata and crossed my fingers. It worked for us, though I do not know the implications here. After doing this surgery, you might try a restore of the past few backup dates and make sure it is successful. Worst case, you rm-rf rdiff-backup-data, though you will loose your increments. If you come up with any good ideas, please let us know! -Eric On Wed, 7 Nov 2007, Kurt Yoder wrote:
Anyone? I'm starting to get worried, since my backups haven't run for several days now. Is there any way to fix the broken rdiff-backup-data directory?On Nov 1, 2007, at 10:26 PM, Kurt Yoder wrote:Hello listI recently started seeing this error on one of my backups. After two backups, it's still appearing:Traceback (most recent call last): File "/usr/bin/rdiff-backup", line 23, in ? rdiff_backup.Main.error_check_Main(sys.argv[1:])File "/var/lib/python-support/python2.4/rdiff_backup/Main.py", line 295, in error_check_Maintry: Main(arglist)File "/var/lib/python-support/python2.4/rdiff_backup/Main.py", line 315, in Maintake_action(rps)File "/var/lib/python-support/python2.4/rdiff_backup/Main.py", line 271, in take_actionelif action == "backup": Backup(rps[0], rps[1])File "/var/lib/python-support/python2.4/rdiff_backup/Main.py", line 328, in Backupbackup_final_init(rpout)File "/var/lib/python-support/python2.4/rdiff_backup/Main.py", line 425, in backup_final_initcheckdest_if_necessary(rpout)File "/var/lib/python-support/python2.4/rdiff_backup/Main.py", line 824, in checkdest_if_necessarydest_rp.conn.regress.Regress(dest_rp)File "/var/lib/python-support/python2.4/rdiff_backup/regress.py", line 71, in Regressfor rf in iterate_meta_rfs(mirror_rp, inc_rpath): ITR(rf.index, rf)File "/var/lib/python-support/python2.4/rdiff_backup/rorpiter.py", line 281, in __call__last_branch.fast_process(*args)File "/var/lib/python-support/python2.4/rdiff_backup/regress.py", line 267, in fast_processif rf.metadata_rorp.isreg(): self.restore_orig_regfile(rf)File "/var/lib/python-support/python2.4/rdiff_backup/regress.py", line 289, in restore_orig_regfiletf.write_from_fileobj(rf.get_restore_fp())File "/var/lib/python-support/python2.4/rdiff_backup/restore.py", line 485, in get_restore_fpreturn robust.check_common_error(error_handler, get_fp)File "/var/lib/python-support/python2.4/rdiff_backup/robust.py", line 32, in check_common_errortry: return function(*args)File "/var/lib/python-support/python2.4/rdiff_backup/restore.py", line 465, in get_fpRdiff.write_patched_fp(current_fp, delta_fp, new_fp)File "/var/lib/python-support/python2.4/rdiff_backup/Rdiff.py", line 73, in write_patched_fprpath.copyfileobj(librsync.PatchedFile(basis_fp, delta_fp), out_fp)File "/var/lib/python-support/python2.4/rdiff_backup/rpath.py", line 58, in copyfileobjinbuf = inputfp.read(blocksize)File "/var/lib/python-support/python2.4/rdiff_backup/librsync.py", line 77, in readself._add_to_outbuf_once()File "/var/lib/python-support/python2.4/rdiff_backup/librsync.py", line 86, in _add_to_outbuf_onceif not self.infile_eof: self._add_to_inbuf()File "/var/lib/python-support/python2.4/rdiff_backup/librsync.py", line 96, in _add_to_inbufnew_in = self.infile.read(blocksize) File "/usr/lib/python2.4/gzip.py", line 225, in read self._read(readsize) File "/usr/lib/python2.4/gzip.py", line 290, in _read self._read_eof() File "/usr/lib/python2.4/gzip.py", line 309, in _read_eof raise IOError, "CRC check failed" IOError: CRC check failedSo my backups are failing. After some quick Google searching, I turned up some discussion about the same kind of error on 1.1.2. It is here:http://osdir.com/ml/sysutils.backup.rdiff-backup.general/2006-03/msg00031.htmlThe only solution seemed to be to delete the rdiff-backup-data directory. I am now using 1.1.5 (Debian Etch, Linux 2.6.18). Was there any other method besides deleting the rdiff directory to fix this?Thanks! _______________________________________________ rdiff-backup-users mailing list at address@hidden http://lists.nongnu.org/mailman/listinfo/rdiff-backup-usersWiki URL: http://rdiff-backup.solutionsfirst.com.au/index.php/RdiffBackupWiki_______________________________________________ rdiff-backup-users mailing list at address@hidden http://lists.nongnu.org/mailman/listinfo/rdiff-backup-users Wiki URL: http://rdiff-backup.solutionsfirst.com.au/index.php/RdiffBackupWiki
[Prev in Thread] | Current Thread | [Next in Thread] |