Re: [BackupPC-users] An idea to fix both SIGPIPE and memory issues with rsync
On Tue, Dec 15, 2009 at 05:42:55PM -0800, Robin Lee Powell wrote: > Just to give a sense of scale here: > > # date ; find /pictures -xdev -type f -printf "%h\n" >/tmp/dirs ; date > Tue Dec 15 12:50:57 PST 2009 > Tue Dec 15 17:26:44 PST 2009 > > (something I ran to try to figure out how to partition the tree) > > The other one isn't even close to finishing, as far as I can tell. Just for amusement: # date ; find /data -xdev -type f -printf "%h\n" >/tmp/dirs ; date Tue Dec 15 12:51:17 PST 2009 [snip many errors caused by things disappearing during the walk] Wed Dec 16 00:55:52 PST 2009 RedHat GFS *really* doesn't like directories with large numbers of files. It's not a big fan of stat() calls, either. -Robin -- They say: "The first AIs will be built by the military as weapons." And I'm thinking: "Does it even occur to you to try for something other than the default outcome?" See http://shrunklink.com/cdiz http://www.digitalkingdom.org/~rlpowell/ *** http://www.lojban.org/ -- This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev ___ BackupPC-users mailing list BackupPC-users@lists.sourceforge.net List:https://lists.sourceforge.net/lists/listinfo/backuppc-users Wiki:http://backuppc.wiki.sourceforge.net Project: http://backuppc.sourceforge.net/
Re: [BackupPC-users] An idea to fix both SIGPIPE and memory issues with rsync
On 12/15 05:42 , Robin Lee Powell wrote: > The other one isn't even close to finishing, as far as I can tell. > In the face of it taking nigh-on 5 hours just to *walk the tree*, > from the local host, I haven't been focusing on little things like > ssh encryption choices too much. :) So you're convinced the speed bottleneck is firmly on the filesystem/disk array? I've noticed that changing the SSH cipher helps substantially sometimes; but sometimes (often?) not much. It's a worthwhile optimization when it works tho. I think I've seen 10-15% improvement in backup times. As for your link issues; the suggestion that it may be a keepalive problem reminded me of this tool: http://freshmeat.net/projects/openssh-watchdog It may be worth trying that. -- Carl Soderstrom Systems Administrator Real-Time Enterprises www.real-time.com -- This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev ___ BackupPC-users mailing list BackupPC-users@lists.sourceforge.net List:https://lists.sourceforge.net/lists/listinfo/backuppc-users Wiki:http://backuppc.wiki.sourceforge.net Project: http://backuppc.sourceforge.net/
Re: [BackupPC-users] An idea to fix both SIGPIPE and memory issues with rsync
On 12/16 06:38 , Carl Wilhelm Soderstrom wrote: > On 12/15 05:42 , Robin Lee Powell wrote: > > The other one isn't even close to finishing, as far as I can tell. > > In the face of it taking nigh-on 5 hours just to *walk the tree*, > > from the local host, I haven't been focusing on little things like > > ssh encryption choices too much. :) > > So you're convinced the speed bottleneck is firmly on the filesystem/disk > array? > > I've noticed that changing the SSH cipher helps substantially sometimes; but > sometimes (often?) not much. It's a worthwhile optimization when it works > tho. I think I've seen 10-15% improvement in backup times. > > As for your link issues; the suggestion that it may be a keepalive problem > reminded me of this tool: > http://freshmeat.net/projects/openssh-watchdog > > It may be worth trying that. According to the main page for the openssh-watchdog: http://www.sc.isc.tohoku.ac.jp/~hgot/sources/openssh-watchdog.html It looks like Debian has a patch that provides this functionality already. Down at the bottom of the page it lists other tools with similar functionality, including: "OpenSSH package from Debian (`ProtocolKeepAlive' option)" -- Carl Soderstrom Systems Administrator Real-Time Enterprises www.real-time.com -- This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev ___ BackupPC-users mailing list BackupPC-users@lists.sourceforge.net List:https://lists.sourceforge.net/lists/listinfo/backuppc-users Wiki:http://backuppc.wiki.sourceforge.net Project: http://backuppc.sourceforge.net/
Re: [BackupPC-users] An idea to fix both SIGPIPE and memory issues with rsync
Robin Lee Powell wrote: > On Tue, Dec 15, 2009 at 05:42:55PM -0800, Robin Lee Powell wrote: >> Just to give a sense of scale here: >> >> # date ; find /pictures -xdev -type f -printf "%h\n" >/tmp/dirs ; date >> Tue Dec 15 12:50:57 PST 2009 >> Tue Dec 15 17:26:44 PST 2009 >> >> (something I ran to try to figure out how to partition the tree) >> >> The other one isn't even close to finishing, as far as I can tell. > > Just for amusement: > > # date ; find /data -xdev -type f -printf "%h\n" >/tmp/dirs ; date > Tue Dec 15 12:51:17 PST 2009 > [snip many errors caused by things disappearing during the walk] > Wed Dec 16 00:55:52 PST 2009 > > RedHat GFS *really* doesn't like directories with large numbers of > files. It's not a big fan of stat() calls, either. Usually it is just simple physics. The disk head can only be in on place at a time. It's never where you want it, and it is slow to move. -- Les Mikesell lesmikes...@gmail.com -- This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev ___ BackupPC-users mailing list BackupPC-users@lists.sourceforge.net List:https://lists.sourceforge.net/lists/listinfo/backuppc-users Wiki:http://backuppc.wiki.sourceforge.net Project: http://backuppc.sourceforge.net/
Re: [BackupPC-users] An idea to fix both SIGPIPE and memory issues with rsync
Robin Lee Powell schrieb: > RedHat GFS *really* doesn't like directories with large numbers of > files. It's not a big fan of stat() calls, either. Well, a network Cluster Filesystem is no fun to backup and might very well be the bottleneck. Ralf -- This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev ___ BackupPC-users mailing list BackupPC-users@lists.sourceforge.net List:https://lists.sourceforge.net/lists/listinfo/backuppc-users Wiki:http://backuppc.wiki.sourceforge.net Project: http://backuppc.sourceforge.net/
Re: [BackupPC-users] An idea to fix both SIGPIPE and memory issues with rsync
I want to start by saying that I appreciate all the help and suggestions y'all have given on something that's obviously not your problem. :) Unfortunately, it looks like this problem is (1) far more interesting than I thought and (2) might be in BackupPC itself. On Tue, Dec 15, 2009 at 11:29:46AM -0900, Chris Robertson wrote: > My guess would be that your firewalls are set up to close > "inactive" TCP sessions. Try adding "-o ServerAliveInterval=60" > to your RsyncClientCmd (so it looks something like "$sshPath -C -q > -x -o ServerAliveInterval=60 -l root $host $rsyncPath $argList+") > and see if that solves your problem. Also suggested by several other people. Well, umm, it kept the connection open. Now I've got [counts] 8 BackupPC_dump processes doing absolutely nothing. The one I was stracing and tcpduming and so on looks fine: ssh is sending the keepalives, and they're being responded to appropriately. The rsync is up on the client, the dump is up on the server, all procs are connected to each other correctly. They're just not *doing* anything. Nothing has errored out; BackupPC thinks everything is fine. Some of the places this is happening are very small backups that usually take a matter of minutes. Suddenly this isn't looking like a networking problem anymore; the networking appears to be just fine. This is looking like a BackupPC problem. My version, by the way, is 3.1.0. It's like at some point BackupPC forgot that it was supposed to be trying to back anything up. Is it possible the nightly jobs are configured in some weird way that broke things? Anyways, the one that has the problem consistently *also* always has it in exactly the same place; I was watching it in basically every way possible, so here comes the debugging stuff. As will become obvious, I probably need to turn the ssh debugging down and the BackupPC debugging up. If, for some reason, you want raw copies of any of this, let me know and I'll post them. On the machine in question, we're backing up /pictures/. /pictures/agate/shared/pictures/ has about a thousand subdirs that each have about a thousand subdirs; the *vast* majority of the files on the file system are in there. The backup always stops doing anything right after lstat of the last file in /pictures/agate/shared/pictures/0216/2987/. According to ls -U, /pictures/agate/shared/pictures/0216/ is the last directory in /pictures/agate/shared/pictures/, and /pictures/agate/shared/pictures/0216/2987/ is the last directory in /pictures/agate/shared/pictures/0216/. Now, there are other directories in /pictures/agate/shared/ that ls -U after pictures/, but with rsync 3.0 on the client, I'm assuming those are handled in parallel. So it looks like the backup is stalling right after finishing checking all the files. No "new" dir is made. There's a NewFileList, but it's empty. XferLog.z is mostly useless; too many -v in my ssh call apparently, but here's the last bit anyways (copied from less): - - browse.jpg^...@^@w<85>@I:1^Kdisplay.jpgp^...@^@y<85>@I:1^Ksegment.jpg<8D>^...@^@t<85>@I:$^Y3187/DSCN1886_gallery.jpg^...@^@}...@i:2^Pfor_cropping.jpg<94>y...@^@<88>^...@i:2^Klibrary.jpg<8A> ^...@^@<8A>^...@i:2^Qmarket_banner.jpgN^...@^@<8C>^...@i:2 browse.jpg~...@^@<80>^...@i:1^D.JPG^T^I^@<8C>^...@i:1 _badge.j...@^@<87>^...@i:2^Lselector.jpg'^...@^@<84>^...@i:3^Nhop_banner.jpg<88>^...@^@<89>^...@i:2^Kdisplay.jpgx...@^@<82>^...@i:2^ksegment.jpgl...@^@^...@i:%(968/food-wrapping-mr-gallery-x_badge.jpgb...@^@^...@i:d^kgallery.jpg^...@^@^...@i:D^Qmarket_banner.jpg^...@^@^...@i:D^Klibrary.jpg^...@^@^...@iD^Oshop_banner.jpg<8C>^...@^@:D browse.jpg&^...@^@@I:C^D.jpg^c...@^^$@I:C^L_segment.jpg ^...@^@@I:d^pfor_cropping.jpg!*...@^z$@I:D^Lselector.jpg<96>^...@^@^...@i:D^Kdisplay.jpgR1^A^@@I:%^K539/061.JPGf...@^@d @I:,^l_gallery.jpgr...@^@a @I:-^Klibrary.jpg^...@^@c @I:-^Pfor_cropping.jpgu...@^@` @I:-^Lselector.jpg<8C>^...@^@b @I:-^Qmarket_banner.jpgl...@^@d @I:-^oshop_banner.jpg^...@^@c @I:- browse.jpg^...@^@b @I:-^Ksegment.jpg<8D>^...@^@a @I:-^kdisplay.j...@^@^...@b @I:-^ibadge.jpg...@^@c @I:$^T7026/blog_browse.jpg^O ^...@^@o...@i:.^Klibrary.jpg^...@^@o...@i:.^lselector.jpg=...@^@o...@i:.^Kgallery.jpg,&^...@^@o...@i:.^Qmarket_banner.jpgl - - The only relevant, and last, LOG line is: 2009-12-16 00:12:00 incr backup started back to 2009-12-12 00:00:08 (backup #0) for directory / Current time on the box is Wed Dec 16 10:54:14 PST 2009 (boxes agree to within a second; didn't test beyond flipping back and forth and typing "date"). Last bit of strace of the dump proc on the SERVER: backuppc 4607 24499 0 00:12 ?00:00:07 /usr/bin/perl /usr/local/bin/BackupPC_dump -i ey02-s00416_pictures - - 04:34:54.081334 read(7, "nner.jpgDX\1\0mW\342I\272-\vlibrary.jpg\177\24\0\0:,\4.JPGN \t\0kW\342I:,\22_market_banner.jpg^\253\1\0lW\342I:$\0376998/P1010001_1_shop_banner.jpg\234\274\0\0\257F\342I:3\4.JPG\25j\t\0\255F\342I:3\f_library.jpg\3
Re: [BackupPC-users] An idea to fix both SIGPIPE and memory issues with rsync
Sorry, a couple of things I forgot. On Wed, Dec 16, 2009 at 11:15:37AM -0800, Robin Lee Powell wrote: > Anyways, the one that has the problem consistently *also* always has > it in exactly the same place; I was watching it in basically every > way possible, so here comes the debugging stuff. As will become > obvious, I probably need to turn the ssh debugging down and the > BackupPC debugging up. I take that back; I've got: $Conf{XferLogLevel} = 2; and there's nothing from BackupPC is the XferLog around or after the time it stopped working. I'm going to set it higher anyways if that's likely to help; what's a good value? Also, no significant CPU activity on either side, so it's not like rsync is sitting there contemplating what files to transfer. Just for interest, memory usage on the client: PID USER PR NI VIRT RES SHR S %CPU %MEMTIME+ COMMAND 21663 root 15 0 436m 402m 640 S0 6.6 198:21.99 /usr/bin/rsync --server --sender ... That's rsync 3.0 *without* --hard-links; a non-trivial burden on a client machine. -Robin -- They say: "The first AIs will be built by the military as weapons." And I'm thinking: "Does it even occur to you to try for something other than the default outcome?" See http://shrunklink.com/cdiz http://www.digitalkingdom.org/~rlpowell/ *** http://www.lojban.org/ -- This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev ___ BackupPC-users mailing list BackupPC-users@lists.sourceforge.net List:https://lists.sourceforge.net/lists/listinfo/backuppc-users Wiki:http://backuppc.wiki.sourceforge.net Project: http://backuppc.sourceforge.net/
Re: [BackupPC-users] An idea to fix both SIGPIPE and memory issues with rsync
On last thing: Stop/Dequeue Backup did the right thing: all parts of the backup, on both server and client, were torn down correctly. So, clearly, there was nothing wrong with the communication between parts as such. -Robin -- They say: "The first AIs will be built by the military as weapons." And I'm thinking: "Does it even occur to you to try for something other than the default outcome?" See http://shrunklink.com/cdiz http://www.digitalkingdom.org/~rlpowell/ *** http://www.lojban.org/ -- This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev ___ BackupPC-users mailing list BackupPC-users@lists.sourceforge.net List:https://lists.sourceforge.net/lists/listinfo/backuppc-users Wiki:http://backuppc.wiki.sourceforge.net Project: http://backuppc.sourceforge.net/
Re: [BackupPC-users] pools are showing data but clients have no backups
Hi, Added some debugging code to Storage/Text.pm sub TextFileWrite, it used to look like this : rename("$file.new", "$file") if ( -f "$file.new" ); I changed it to : if ( -f "$file.new" ) { my $renRet = rename("$file.new", "$file"); if ($renRet) { print "Rename OK\n"; } else { print "Rename failed $file.new to $file\n"; } } else { print "failed -f $file.new\n"; } This is what BackupPC_fixupBackupSummary returns when I run it as root for one of the backup hosts where backups has only up to #45 and backups.new has #46: Reading /glfsdist/backuppc1/pc/home_caox/46/backupInfo Adding info for backup 46 from backupInfo file Rename failed /glfsdist/backuppc1/pc/home_caox/backups.new to /glfsdist/backuppc1/pc/home_caox/backups [r...@gluster1 ~]# ls -l /glfsdist/backuppc1/pc/home_caox/backups* -rw-r- 1 root root 3440 Dec 15 12:41 /glfsdist/backuppc1/pc/home_caox/backups -rw-r- 1 root root 3514 Dec 16 14:00 /glfsdist/backuppc1/pc/home_caox/backups.new Let's try a manual move: [r...@gluster1 ~]# mv /glfsdist/backuppc1/pc/home_caox/backups.new /glfsdist/backuppc1/pc/home_caox/backups mv: overwrite `/glfsdist/backuppc1/pc/home_caox/backups'? y mv: cannot move `/glfsdist/backuppc1/pc/home_caox/backups.new' to `/glfsdist/backuppc1/pc/home_caox/backups': File exists hrmm... here's the strace -ff : stat("/glfsdist/backuppc1/pc/home_caox/backups", y ^ that's me hitting 'y' when it asks if I want to overwrite the file. {st_mode=S_IFREG|0640, st_size=3440, ...}) = 0 lstat("/glfsdist/backuppc1/pc/home_caox/backups.new", {st_mode=S_IFREG|0640, st_size=3514, ...}) = 0 lstat("/glfsdist/backuppc1/pc/home_caox/backups", {st_mode=S_IFREG|0640, st_size=3440, ...}) = 0 stat("/glfsdist/backuppc1/pc/home_caox/backups", {st_mode=S_IFREG|0640, st_size=3440, ...}) = 0 geteuid() = 0 getegid() = 0 getuid()= 0 getgid()= 0 access("/glfsdist/backuppc1/pc/home_caox/backups", W_OK) = 0 rename("/glfsdist/backuppc1/pc/home_caox/backups.new", "/glfsdist/backuppc1/pc/home_caox/backups") = -1 EEXIST (File exists) open("/usr/share/locale/locale.alias", O_RDONLY) = 3 fstat(3, {st_mode=S_IFREG|0644, st_size=2528, ...}) = 0 mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x2b6fc324b000 read(3, "# Locale name alias data base.\n#"..., 4096) = 2528 read(3, "", 4096) = 0 close(3)= 0 munmap(0x2b6fc324b000, 4096)= 0 open("/usr/share/locale/en_US.UTF-8/LC_MESSAGES/coreutils.mo", O_RDONLY) = -1 ENOENT (No such file or directory) open("/usr/share/locale/en_US.utf8/LC_MESSAGES/coreutils.mo", O_RDONLY) = -1 ENOENT (No such file or directory) open("/usr/share/locale/en_US/LC_MESSAGES/coreutils.mo", O_RDONLY) = -1 ENOENT (No such file or directory) open("/usr/share/locale/en.UTF-8/LC_MESSAGES/coreutils.mo", O_RDONLY) = -1 ENOENT (No such file or directory) open("/usr/share/locale/en.utf8/LC_MESSAGES/coreutils.mo", O_RDONLY) = -1 ENOENT (No such file or directory) open("/usr/share/locale/en/LC_MESSAGES/coreutils.mo", O_RDONLY) = -1 ENOENT (No such file or directory) write(2, "mv: ", 4mv: ) = 4 write(2, "cannot move `/glfsdist/backuppc1"..., 104cannot move `/glfsdist/backuppc1/pc/home_caox/backups.new' to `/glfsdist/backuppc1/pc/home_caox/backups') = 104 open("/usr/share/locale/en_US.UTF-8/LC_MESSAGES/libc.mo", O_RDONLY) = -1 ENOENT (No such file or directory) open("/usr/share/locale/en_US.utf8/LC_MESSAGES/libc.mo", O_RDONLY) = -1 ENOENT (No such file or directory) open("/usr/share/locale/en_US/LC_MESSAGES/libc.mo", O_RDONLY) = -1 ENOENT (No such file or directory) open("/usr/share/locale/en.UTF-8/LC_MESSAGES/libc.mo", O_RDONLY) = -1 ENOENT (No such file or directory) open("/usr/share/locale/en.utf8/LC_MESSAGES/libc.mo", O_RDONLY) = -1 ENOENT (No such file or directory) open("/usr/share/locale/en/LC_MESSAGES/libc.mo", O_RDONLY) = -1 ENOENT (No such file or directory) write(2, ": File exists", 13: File exists) = 13 write(2, "\n", 1 ) = 1 close(1)= 0 exit_group(1) = ? This has got to be some glusterfs weirdness but why? any ideas? Thanks, Sabuj > The file backups.new is written out by a function called BackupInfoWrite > (which further uses a function called TextFileWrite), which is called > both in BackupPC_dump, when a backup completes and in BackupPC_link, > when the linking of the backup completes. The file backups.new is > written, verified and then renamed to backups. If the verification > fails, the rename does not occur. I'd suggest checking for permissions > issues or file system corruption. > Chris --
Re: [BackupPC-users] Backuppc only works from the command line
M. Sabath wrote: > Hello all, > > I use backuppc on Debian 5. > Since I upgraded from Debian 4 to Debian 5 backuppc doesn't run > automatically. > > Our server runs only during daytime between 7am and 19 pm > Let me see if I have this right... Your server is only powered on from 7 am to 7 pm... > >From the command line all works fine. > > Using backuppc with Debian 4.0 all worked fine. > > What am I doing wrong? > > > Thank you > > Markus > > > -- > > > Here are some configuration entries of my config.pl which might be > interesting: > > $Conf{WakeupSchedule} = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, > 15, 16, 17, 18, 19, 20, 21, 22, 23]; > > > $Conf{FullPeriod} = 6.97; > $Conf{IncrPeriod} = 0.97; > > $Conf{FullKeepCnt} = [3,1,1]; > > $Conf{BackupsDisable} = 0; > > $Conf{BlackoutPeriods} = [ > { > hourBegin => 7.0, > hourEnd => 19.5, > ...your blackout period covers that whole time... > weekDays => [1, 2, 3, 4, 5], > ...at least on the week days, and you are wondering why BackupPC is not working automatically? > }, > ]; > Let me know if my understanding is not correct. Otherwise, I'd suggest reading the fine manual: http://backuppc.sourceforge.net/faq/BackupPC.html#item__conf_blackoutperiods_ Chris -- This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev ___ BackupPC-users mailing list BackupPC-users@lists.sourceforge.net List:https://lists.sourceforge.net/lists/listinfo/backuppc-users Wiki:http://backuppc.wiki.sourceforge.net Project: http://backuppc.sourceforge.net/
[BackupPC-users] Slow link options
I have a few remote sites I am wanting to backup using backuppc. However, two are on slow DSL connections and the other 2 are on T1's. I did some math and roughly figured that the DSL connections, having a 256k upload, could do approximately 108MB/hour of transfer. With these clients having around 65GB each that would take FOREVER!!! I am able to take the backuppc server to 2 of the remote locations (the DSL ones) and put it on the LAN with the server to be backed up to get the initial full backup. What I am wondering is this: What do others do with slow links like this? I need a full backup at least weekly and incrimentals nightly. Is there an easy way around this? Thanks in advance. -- This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev ___ BackupPC-users mailing list BackupPC-users@lists.sourceforge.net List:https://lists.sourceforge.net/lists/listinfo/backuppc-users Wiki:http://backuppc.wiki.sourceforge.net Project: http://backuppc.sourceforge.net/
Re: [BackupPC-users] An idea to fix both SIGPIPE and memory issues with rsync
Robin Lee Powell wrote: > > They're just not *doing* anything. Nothing has errored out; BackupPC > thinks everything is fine. > > Some of the places this is happening are very small backups that > usually take a matter of minutes. > > Suddenly this isn't looking like a networking problem anymore; the > networking appears to be just fine. This is looking like a BackupPC > problem. My version, by the way, is 3.1.0. It's like at some point > BackupPC forgot that it was supposed to be trying to back anything > up. This sounds vaguely familiar, perhaps like: http://www.mail-archive.com/backuppc-de...@lists.sourceforge.net/msg00321.html but I don't know if that had a resolution. It doesn't seem common, though. > The backup always stops doing anything right after lstat of the last > file in /pictures/agate/shared/pictures/0216/2987/. According to ls > -U, /pictures/agate/shared/pictures/0216/ is the last directory in > /pictures/agate/shared/pictures/, and > /pictures/agate/shared/pictures/0216/2987/ is the last directory in > /pictures/agate/shared/pictures/0216/. Now, there are other > directories in /pictures/agate/shared/ that ls -U after pictures/, > but with rsync 3.0 on the client, I'm assuming those are handled in > parallel. I don't think that's true. Backuppc is going to insist on protocol 28 which, I think, means it has to get the entire directory tree before starting the comparison. > I'm quite stumped. Any ideas? Is your client rsync binary up to date with any distribution updates? Maybe it is just a bug that has already been fixed. Also, I thought at one point you said this was only happening with incrementals. Have you tried forcing a full run? -- Les Mikesell lesmikes...@gmail.com -- This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev ___ BackupPC-users mailing list BackupPC-users@lists.sourceforge.net List:https://lists.sourceforge.net/lists/listinfo/backuppc-users Wiki:http://backuppc.wiki.sourceforge.net Project: http://backuppc.sourceforge.net/
Re: [BackupPC-users] Slow link options
Kameleon wrote: > I have a few remote sites I am wanting to backup using backuppc. > However, two are on slow DSL connections and the other 2 are on T1's. > I did some math and roughly figured that the DSL connections, having a > 256k upload, could do approximately 108MB/hour of transfer. With these > clients having around 65GB each that would take FOREVER!!! > > I am able to take the backuppc server to 2 of the remote locations > (the DSL ones) and put it on the LAN with the server to be backed up > to get the initial full backup. What I am wondering is this: What do > others do with slow links like this? I need a full backup at least > weekly and incrimentals nightly. Is there an easy way around this? The feasibility of this depends entirely on the rate of change of the backup data. Once you get the initial full, rsync backups only transfer changes. Have a look at the documentation (http://backuppc.sourceforge.net/faq/BackupPC.html#backup_basics) for more details. > > Thanks in advance. Chris -- This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev ___ BackupPC-users mailing list BackupPC-users@lists.sourceforge.net List:https://lists.sourceforge.net/lists/listinfo/backuppc-users Wiki:http://backuppc.wiki.sourceforge.net Project: http://backuppc.sourceforge.net/
Re: [BackupPC-users] An idea to fix both SIGPIPE and memory issues with rsync
On Wed, Dec 16, 2009 at 03:13:23PM -0600, Les Mikesell wrote: > Robin Lee Powell wrote: > > > > They're just not *doing* anything. Nothing has errored out; BackupPC > > thinks everything is fine. > > > > Some of the places this is happening are very small backups that > > usually take a matter of minutes. > > > > Suddenly this isn't looking like a networking problem anymore; the > > networking appears to be just fine. This is looking like a BackupPC > > problem. My version, by the way, is 3.1.0. It's like at some point > > BackupPC forgot that it was supposed to be trying to back anything > > up. > > This sounds vaguely familiar, perhaps like: > http://www.mail-archive.com/backuppc-de...@lists.sourceforge.net/msg00321.html > but I don't know if that had a resolution. It doesn't seem common, though. That part turns out to be Extremely Lame: it was adding "-vv" to ssh. Drop the -vs and put the -q back in, and that part gets better. Which means I'm sort of back to the drawing board. I'll try with ServerAliveInterval=60 and -q and let y'all know. > > The backup always stops doing anything right after lstat of the > > last file in /pictures/agate/shared/pictures/0216/2987/. > > According to ls -U, /pictures/agate/shared/pictures/0216/ is the > > last directory in /pictures/agate/shared/pictures/, and > > /pictures/agate/shared/pictures/0216/2987/ is the last directory > > in /pictures/agate/shared/pictures/0216/. Now, there are other > > directories in /pictures/agate/shared/ that ls -U after > > pictures/, but with rsync 3.0 on the client, I'm assuming those > > are handled in parallel. > > I don't think that's true. Backuppc is going to insist on > protocol 28 which, I think, means it has to get the entire > directory tree before starting the comparison. Then it's even weirder. :) But see above. -Robin -- They say: "The first AIs will be built by the military as weapons." And I'm thinking: "Does it even occur to you to try for something other than the default outcome?" See http://shrunklink.com/cdiz http://www.digitalkingdom.org/~rlpowell/ *** http://www.lojban.org/ -- This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev ___ BackupPC-users mailing list BackupPC-users@lists.sourceforge.net List:https://lists.sourceforge.net/lists/listinfo/backuppc-users Wiki:http://backuppc.wiki.sourceforge.net Project: http://backuppc.sourceforge.net/
[BackupPC-users] problem with smb transfers
so i've set up my backuppc server again (we got a bunch of new drives and new server) and rather than upgrade, i just did a fresh installation. so i've migrated my settings from v2.1 to v3.2.0beta0, and i'm having a problem with smb transfers. they seem to fail. here's a sample xfer error log file: Running: /usr/bin/smbclient prettylady\\c\$ -I 192.168.0.244 -U backupuser -E -N -d 1 -c tarmode\ full -Tc - \\_data_\\\* \\faxserver\\\* full backup started for share c$ Xfer PIDs are now 20691,20690 Anonymous login successful [ skipped 1 lines ] tree connect failed: NT_STATUS_ACCESS_DENIED Anonymous login successful [ skipped 1 lines ] tree connect failed: NT_STATUS_ACCESS_DENIED tarExtract: Done: 0 errors, 0 filesExist, 0 sizeExist, 0 sizeExistComp, 0 filesTotal, 0 sizeTotal Got fatal error during xfer (No files dumped for share c$) Backup aborted (No files dumped for share c$) Not saving this as a partial backup since it has fewer files than the prior one (got 0 and 0 files versus 0) what i don't get is why is it trying to log in anonymously?? it should be logging in as "backupuser", and the smbclient seems to indicate that. the config file for this machine is: $Conf{SmbShareName} = 'c$'; $Conf{BackupFilesOnly} = { 'c$' => [ '\_data_\*', '\faxserver\*' ], }; and in the config.pl, the lines: $Conf{SmbShareUserName} = 'backupuser'; $Conf{SmbSharePasswd} = ''; exist. when i try to log into the host from another windows machine, to the c$ share, using the user backupuser and the password, it connects fine. i suspect i've forgotten to update a line in my config.pl. (i didn't copy the one from v2.1 over, i did a fresh installation, i'm starting with a fresh pool, and so i went through the config.pl and set it all up again. what did i forget? thanks! -- This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev ___ BackupPC-users mailing list BackupPC-users@lists.sourceforge.net List:https://lists.sourceforge.net/lists/listinfo/backuppc-users Wiki:http://backuppc.wiki.sourceforge.net Project: http://backuppc.sourceforge.net/
[BackupPC-users] problem with smb transfers (solved)
(unfortunately i don't receive copies of my own emails, but...) i figured out what was going on. (or google did... of course...) from the page: https://bugs.launchpad.net/ubuntu/+source/backuppc/+bug/283652 "I believe that the fix you need is to edit /etc/backuppc/config.pl. There are three strings: $Conf{SmbClientFullCmd} $Conf{SmbClientIncrCmd}, and $Conf{SmbClientRestoreCmd} which control Samba backups and restore. In all three strings remove the "-N" flag. My understanding that the flag is no longer needed, because the login prompt is automatically suppressed because backuppc passes the password through the PASSWD environment variable. I reported separately a bug ( https://bugs.launchpad.net/bugs/297025) in the new version of smbclient that prevents this password from being passed if the -N flag is used. Let me know if this doesn't work." but just thought i'd report back!! -- This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev ___ BackupPC-users mailing list BackupPC-users@lists.sourceforge.net List:https://lists.sourceforge.net/lists/listinfo/backuppc-users Wiki:http://backuppc.wiki.sourceforge.net Project: http://backuppc.sourceforge.net/
Re: [BackupPC-users] Backuppc only works from the command line
Am Mittwoch, den 16.12.2009, 11:27 -0900 schrieb Chris Robertson: > M. Sabath wrote: > > Hello all, > > > > I use backuppc on Debian 5. > > Since I upgraded from Debian 4 to Debian 5 backuppc doesn't run > > automatically. > > > > Our server runs only during daytime between 7am and 19 pm > > > > Let me see if I have this right... Your server is only powered on from > 7 am to 7 pm... > > > >From the command line all works fine. > > > > Using backuppc with Debian 4.0 all worked fine. > > > > What am I doing wrong? > > > > > > Thank you > > > > Markus > > > > > > -- > > > > > > Here are some configuration entries of my config.pl which might be > > interesting: > > > > $Conf{WakeupSchedule} = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, > > 15, 16, 17, 18, 19, 20, 21, 22, 23]; > > > > > > $Conf{FullPeriod} = 6.97; > > $Conf{IncrPeriod} = 0.97; > > > > $Conf{FullKeepCnt} = [3,1,1]; > > > > $Conf{BackupsDisable} = 0; > > > > $Conf{BlackoutPeriods} = [ > > { > > hourBegin => 7.0, > > hourEnd => 19.5, > > > > ...your blackout period covers that whole time... > > > weekDays => [1, 2, 3, 4, 5], > > > > ...at least on the week days, and you are wondering why BackupPC is not > working automatically? > > > }, > > ]; > > > > Let me know if my understanding is not correct. Otherwise, I'd suggest > reading the fine manual: > http://backuppc.sourceforge.net/faq/BackupPC.html#item__conf_blackoutperiods_ > > Chris > Hello Chris, yes you are rigth. So i have to disable the blackout period. Is this possible? Otherwise I will set it to a time where our server is off. Thank you MArkus -- This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev ___ BackupPC-users mailing list BackupPC-users@lists.sourceforge.net List:https://lists.sourceforge.net/lists/listinfo/backuppc-users Wiki:http://backuppc.wiki.sourceforge.net Project: http://backuppc.sourceforge.net/