Re: Overlapping backups: should I expect problems?
On Mon, Sep 15, 2008 at 11:36 PM, John Morris <[EMAIL PROTECTED]> wrote: > My question is, is this actually a good idea? Will there be any problems I > haven't anticipated, such as the two configs conflicting? For example, I > notice that there is only one /etc/amandates file that is presumably shared > by both configurations. I don't know whether the date to begin incrementals > from is from this file or from /var/lib/amanda/gnutar-lists. No, it isn't a good idea, because there's a better solution that doesn't involve this sort of gymnastics: RAIT. You can set up a RAIT between the two tape devices, and then even if a drive is down, your recoveries will work fine. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: amrecover with 2.6.0p2
On Mon, Sep 15, 2008 at 4:55 PM, Matt Burkhardt <[EMAIL PROTECTED]> wrote: > I'm trying to run amrecover to get back some files, and I'm getting > > amrecover: error while loading shared libraries: libamclient-2.6.0p1.so: > cannot open shared object file: No such file or directory > > and the only files I have on my box are the p2.so files - so do I need to > recompile, reuse, something? Sounds like a recompile -- at least of amrecover. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: planner: could not lock log file
On Fri, Sep 12, 2008 at 11:28 AM, John Hein <[EMAIL PROTECTED]> wrote: > I think amanda has required gmake for quite a while, hasn't it? Quietly, yes -- some GNU extensions snuck in/around 2.5.0 or 2.5.1, and nobody complained, so when I rewrote the autoconf/automake stuff, I just went whole-hog to using GNU make pattern substitutions to build all of these perl and shell scripts without a lot of Makefile acrobatics. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: Single tape changer process
On Thu, Sep 11, 2008 at 11:53 PM, Olivier Nicole <[EMAIL PROTECTED]> wrote: > Is there a mechanism in Amanda to ensure that only a single tape > changer process is running at any given time? No -- and this poses a problem for processes that want to move data between devices. I'm working on such a process, and for that reason I'm working on an overhaul of the changer API at the moment. The key problem with the existing API is that it has no way to indicate that a process is finished with a device, and that the changer can load a new volume into that device. > Is there any mechanism in Amanda that prevents two processes that > access tape to be running at same time (dump, flush, restore, > recover)? The "$logdir/log" file gets locked via amflock by any process that accesses the tape drive. This is much more exclusive than necessary, but it is effective. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: planner: could not lock log file
On Fri, Sep 12, 2008 at 7:58 AM, Taalaibek Ashirov <[EMAIL PROTECTED]> wrote: > Yes, in the test the problem goes away. Actually, I've installed amanda > from ports. Now I deinstalled amanda and got latest source version from > amandas website, added those lines to amflock(). When I did make it gave > me error like "lock is not declared". I declared lock same as test > (struct flock lock;). This time I got this: The problem you encountered is, I think, already fixed in 2.6.0p2. So you shouldn't have to apply any patches to the source before compiling. > creating amflock-test > make: don't know how to make amgpgcrypt. Stop > *** Error code 2 > Stop in /root/amanda-2.6.0p2/common-src. > *** Error code 1 > Stop in /root/amanda-2.6.0p2. > web# Amanda requires GNU make now (gmake). Dustin -- Storage Software Engineer http://www.zmanda.com
Re: [Amanda-users] Amanda 2.6.0p2 + S3 results in cURL errors
On Thu, Sep 11, 2008 at 2:12 PM, moekyle <[EMAIL PROTECTED]> wrote:
> Here is what I posted to the bug tracker in sourceforge.
Oh, dear -- the sourceforge bug tracker is completely unused these
days. If there are others on the list who have submitted things to
the sourceforge tracker, please post to the list. I should find a way
to hide that tracker.
> delete_file(S3Device *self,
> int file)
> {
> gboolean result;
> GSList *keys;
> char *my_prefix = g_strdup_printf("%sf%08x-", self->prefix, file);
>
> result = s3_list_keys(self->s3, self->bucket, self->prefix, "-",
> &keys);
> if (!result) {
[snip]
> Not sure whay the my_prefix is trying to do but by removing it from the
> s3_list_keys it fixed my issue and now backups work correctly.
> result = s3_list_keys(self->s3, self->bucket, self->prefix, "-",
> &keys);
If you look at the documentation for s3_list_keys, you'll see that it
lists all strings matching PREFIX*DELIMITER*, but only including the
PREFIX*DELIMITER portion. The S3 device names objects in a bucket
like
slot-01f001e-filestart
..
slot-01f001eb03ad.data
slot-01f001eb03ae.data
slot-01f001eb03af.data
..
The first object is the header, and the remainder are blocks of data,
where the 'b' in the middle is the border between the file number
(${FILENUM}, 0x1e in this case) and the block number within that file
(0x3ad..0x3ae shown here). 'slot-01' here is the device prefix
(${DPFX}). The pattern as present in the released source code looks
for all objects matching "${DPFX}f${FILENUM}*", and requests the full
name of each. It then proceeds to delete each of those objects -- in
effect, deleting all data with the given file number.
With your patch, you're asking for all files matching "${DPFX}*-*",
returning only the portion matching "${DPFX}*-". This will match the
"tapestart" object for *all* files, but only return the first part of
the object name:
slot-01f0001-
slot-01f0002-
slot-01f0003-
..
these keys do not exist, so the deletion should fail -- but perhaps
Amazon does not respond with an error when deleting a nonexistent
object? Anyway, the end result is that you've effectively disabled
delete_file, so you are probably using more S3 storage than you need,
and will get old data intermingled with new data when you try to make
a recovery. This avoids the error Lisa reported simply by masking the
problem.
On looking more deeply, I think I see the problem: s3_list_keys, or in
particular list_fetch, limits the response to 100k. I'll need to dig
into this a little more deeply, but I should have a patch out shortly.
--
Storage Software Engineer
http://www.zmanda.com
Re: Amanda 2.6.0p2 + S3 results in cURL errors
On Wed, Sep 10, 2008 at 4:10 PM, Lisa Seelye <[EMAIL PROTECTED]> wrote: > While listing S3 keys: CURL error: Failed writing body (CURLcode 23) > taper: Error writing label BUCKETNAME012 to device BUCKETNAME012/. First, I would recommend using a single bucket for all of your Amanda backups, with device names such as s3:BUCKETNAME/tape012 since Amazon limits you to 100 buckets. CURLE_WRITE_ERROR (23) An error occurred when writing received data to a local file, or an error was returned to libcurl from a write callback. My best guess is that Amazon is returning more data than Amanda expects. Can you add device_property "VERBOSE" "YES" to your amanda.conf and re-run to see what additional information appears in the logs? Dustin -- Storage Software Engineer http://www.zmanda.com
Re: planner: could not lock log file
On Wed, Sep 10, 2008 at 8:16 AM, Taalaibek Ashirov <[EMAIL PROTECTED]> wrote: > planner: could not lock log file /var/log/amanda/dotProject/log: Invalid > argument ... > $amadmin x version | grep LOCKING > LOCKING=POSIX_FCNTL DEBUG_CODE AMANDA_DEBUG_DAYS=4 Is there some reason that fcntl-based locking would not work properly on that filesystem? One solution may be to recompile with a different locking mechanism (the lnlock implementation should work). Dustin -- Storage Software Engineer http://www.zmanda.com
Re: backwards compatibility
On Tue, Sep 9, 2008 at 2:05 PM, Krahn, Anderson <[EMAIL PROTECTED]> wrote: > Will a newer Amanda backup server be compatible with older clients. For > instance, I have already installed Amanda 2.6.0p2 on a server and will > possibly need to backup clients that have Amanda 2.5.2p1 from a remote data > center. Yep. See http://wiki.zmanda.com/index.php/Version_compatibility for the full story. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: connection problem during `sendbackup' stage
On Tue, Sep 9, 2008 at 8:50 AM, Jukka Salmi <[EMAIL PROTECTED]> wrote: > c:10080 -> s:846 udp Amanda 2.5 REP HANDLE ... CONNECT DATA 56639 MESG 56638 > INDEX 56637 ... > s:846 -> c:10080 udp Amanda 2.4 ACK HANDLE ... > > s:50029 -> c:56639 tcp SYN > c:56639 -> s:50029 tcp SYN,RST > > Oops. IIUC this TCP connection is supposd to transfer the actual data > to back up. Hmm, why could it be reset by the client? Most likely, the port is closed by the time the server tries to contact it, because something has gone wrong on the client. Note that the index tee uses 'sed'. Is that sed invocation failing on NetBSD? Check the sendbackup debug logs. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: caution: gtar 1.20 & amanda < 2.5.1
On Sun, Sep 7, 2008 at 10:35 AM, John Hein <[EMAIL PROTECTED]> wrote: > Just curious... Why doesn't a wiki search for gnutar or gnu tar find this? The wikimedia search is based on MySQL's full-text search, which isn't very good. In particular, it doesn't index or find words of three or fewer characters. I just added the word "gnutar" to the page, which should help. I also added a link to it from the "User Documentation" page. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: caution: gtar 1.20 & amanda < 2.5.1
On Sun, Sep 7, 2008 at 12:44 AM, John Hein <[EMAIL PROTECTED]> wrote: > There was a zmanda wiki page that described issues with various gnutar > and amanda version combinations, but I can't seem to find it at the > moment (the search doesn't turn it up that I saw). > > If someone finds it, let me know and I'll try to see that this info > gets added. The link is: http://wiki.zmanda.com/index.php/FAQ:What_versions_of_GNU_Tar_are_Amanda-compatible%3F which does specify that those versions will not get along, but not for the reason you described, so it's definitely worth an edit. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: Move disk without loosing backup history?
On Thu, Aug 28, 2008 at 7:48 AM, Gene Heskett <[EMAIL PROTECTED]> wrote: > The tar folks at gnu do not consider that a bug, but as a part of the > security. To be fair, the tar developers did fix this -- in 1.21, IIRC. As to the original question, no, I don't think that there is any way to "pretend" that one host/disk is the same as another. When we have a more expressive catalog, it might be interesting to have some kind of "redirect" that amrecover could follow when browsing the history of a DLE -- similar to Subversion's "copy-from". Dustin -- Storage Software Engineer http://www.zmanda.com
Re: tar ignoring --one-file-system
Have a look at the runtar debug log to see exactly how it's invoking tar, just to be sure. Do your incremental files specify directories that cross devices? Does Amanda cross devices even when doing a level 0? Dustin -- Storage Software Engineer http://www.zmanda.com
Re: What tapedev is used by Amanda
On Thu, Aug 21, 2008 at 10:57 PM, Olivier Nicole <[EMAIL PROTECTED]> wrote: > Does Amanda use the tapedev defined in amanda.conf or the tapedev > returned by the changer? That returned by the changer. What makes it confusing is that some changers return the tapedev defined in amanda.conf :) > If Amanda uses the tapedev returned by the changer, I think that > chg-disk could be rewritten to avoid using symlinks, returning the > directory of the slot each time, and so could work with file systems > that does not implement symlinks (NTFS/FAT32 USB disks). Yep, sounds like a good plan. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: multiple amdumps
(please reply to the list) On Thu, Aug 21, 2008 at 6:50 PM, aminukapon <[EMAIL PROTECTED]> wrote: > I have 2 main directories needed to be backed up and each has several sub > directoires each sub directories should be easily recovered if I used > different Dailysets. You should just add each directory as a separate DLE. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: multiple amdumps
On Thu, Aug 21, 2008 at 1:33 PM, aminukapon <[EMAIL PROTECTED]> wrote: > I would like to find out if i could run amdump on several DailySet at the > same time. If it is possible ,how should I go about doing this I wouldn't recommend it. Why do you want to do this? Dustin -- Storage Software Engineer http://www.zmanda.com
Re: Amanda on Macos X
On Thu, Aug 21, 2008 at 4:38 AM, Nouveaux Territoires <[EMAIL PROTECTED]> wrote: > Ok for the authentication but how do you run the amandad server ? > Like Tiger ? I don't have Leopard yet, so I don't know. You should try it out, and report your success on the wiki, or any problems here. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: Changer policy
On Thu, Aug 21, 2008 at 8:55 AM, Olivier Nicole <[EMAIL PROTECTED]> wrote: > Does it exist a changer that allow the manual change of a caroussel, > so I don't have to redesign everything from scratch? (I mean the > Amanda phylosophy, not the exact way to load/eject a caroussel or > mout/unmout a disk). I *think* that chg-manual will do what you want, perhaps with a little bit of tweaking to recognize when vtapes are on the same disk. This is a fairly common use-case, so if you're working against 2.6.0, I'm sure a lot of other people would like to see your solution. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: Syntax error in chg-disk
On Thu, Aug 21, 2008 at 5:37 AM, Olivier Nicole <[EMAIL PROTECTED]> wrote: > I am using chg-disk on Amanda 2.5.1p3 on FreeBSD and while I was > trying it by hand, I found a Bourne Sheel syntax error: Thanks for the report! By the way, if you have any sway with the FreeBSD folks, please poke them to upgrade their port at some point. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: Using a tape autoloader
On Tue, Aug 19, 2008 at 4:12 PM, Adriana Liendo <[EMAIL PROTECTED]> wrote: > I'm trying to configure Amanda, using a tape autoloader. I think I could use > two tape during the backup. Therefore, I'm trying to set amanda so that it > uses tpchnger "chg-multi". But this device has just 1 driver (/dev/nst0). My > question is that when I configure the file chg-multi.conf I have to set the > slots to be used. I set just 1 slot, but it says that is empty (changer > error: slot 1: chg-multi: slot is empty). I don't know if I'm using a wrong > option for the tapes control. You actually want chg-zd-mtx (http://wiki.zmanda.com/index.php/Chg-zd-mtx). Chg-multi is for a setup where you have no autoloader but multiple drives. Dustin -- Storage Software Engineer http://www.zmanda.com
Glib detection problems -- troubleshooting page
I just added http://wiki.zmanda.com/index.php/Could_not_run_GLIB_test_program%2C_checking_why for the Glib problems we seem to be hearing so much about. Please feel free to add your own suggestions there! Dustin -- Storage Software Engineer http://www.zmanda.com
Re: Glib detection failing
On Mon, Aug 18, 2008 at 12:19 PM, Jon LaBadie <[EMAIL PROTECTED]> wrote: > I've see some gnome software that has funny numbering. > Like 2.1 > 2.11 > 2.12 > 2.2 > 2.21 > So it could look like 2.11 and 2.12 etc are newer than 2.2. > I don't know if your glib 2.16 is newer than 2.2 or not. It is newer, no worries. I'm going to put a FAQ on the wiki about this. The error in config.log is: /usr/local/include/glib-2.0/glib/gmessages.h:128: warning: invalid character in macro parameter name /usr/local/include/glib-2.0/glib/gmessages.h:128: badly punctuated parameter list in `#define' In my gmessages.h, that's 125 #ifdef G_HAVE_ISO_VARARGS 126 #define g_error(...)g_log (G_LOG_DOMAIN, \ 127G_LOG_LEVEL_ERROR,\ 128__VA_ARGS__) which suggests that glib was built with a compiler that understands __VA_ARGS__ (I think this is a C99 thing), but that you're not building Amanda with such a compiler. It looks like configure is finding gcc-2.7.2.1. Amanda itself requires a C99 compiler, so you'll probably need to use a newer gcc. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: Glib detection failing
On Mon, Aug 18, 2008 at 11:10 AM, stan <[EMAIL PROTECTED]> wrote: > I cna sned config.log, if it will help. That would be great. Just a look at it may give some additional information, too. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: zmanda SRPM error
On Thu, Aug 14, 2008 at 11:56 AM, FM <[EMAIL PROTECTED]> wrote:
> Redhat enterprise 5.2 32 bits version
(replying for dan, who's on a plane)
The ".2" confuses the integer expression on line 77. Since the
release, we haved changed to using expressions like this:
%if %(awk '$1 == "Red" && $7 ~ /5.*/ { exit 1; }'
/etc/redhat-release; echo $?)
You can edit the .spec file similarly.
Dustin
--
Storage Software Engineer
http://www.zmanda.com
Re: Amanda talk at CALUG (Columbia, MD)
Apologies for the self-reply, but as a brief report on this evening's presentation. Nikolas Coukouma, our newest Amanda developer, and Jon LaBadie, a seasoned member of the Amanda community, were both present and added valuable comments and perspectives to the conversation. I talked about Amanda's strengths, walked through a very basic Amanda configuration, talked about all the fun new projects in Amanda (Device API, App API, Xfer Architecture, and Perl Rewrite), and then spent some time talking generally about open-source communities and development. Slides are at http://wiki.zmanda.com/images/a/a4/Amanda-calug.pdf (note that some of these slides were held "in reserve" and were not presented tonight). Please have a look. I'd be happy to continue any of these conversations online, for those who were there or whose interest is piqued by the slides. Dustin -- Storage Software Engineer http://www.zmanda.com
Amanda talk at CALUG (Columbia, MD)
If you are in the Washington, DC area, I'm giving a talk at the Columbia Area Linux Users' Group at 7pm tonight. More information is at http://calug.org The last guy who heckled me at a talk is now a full-time Zmanda employee, if that's any incentive! Dustin -- Storage Software Engineer http://www.zmanda.com
Re: 2.6 upgrade file locations
--with-sbindir=/opt/amanda2/sbin \ --with-libexecdir=/opt/amanda2/libexec \ --with-libdir=/opt/amanda2/lib \ --with-mandir=/opt/amanda2/man \ These four lines are being ignored -- thes options should not include "with-". Check the configure --help. > 164 ./perlib/Amanda This has to be present for perl to find packages with the "Amanda::" prefix. > 808 ./share/amanda What's in here? I can't tell just from the directory name where it came from. > 4 ./include/amanda This is amincludedir, which you can set explicitly if you'd like (--with-amincludedir). See configure --help. > 76 ./var/lib/amanda/example/label-templates > 124 ./var/lib/amanda/example > 60 ./var/lib/amanda/template.d > 188 ./var/lib/amanda These currently can't be changed (/var/lib/amanda = $(localstatedir)/amanda. You could fix that with a patch to config/amanda/dirs.m4 and example/Makefile.am. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: very large filecount filesystems
On Mon, Aug 11, 2008 at 1:07 PM, Vectorz Sigma <[EMAIL PROTECTED]> wrote: > I have a mail server that has a very large number of files on its fs. We're > talking in the 10's of millions of files. Therefore, as you can imagine, > the differential comparison calculations nightly take a very long time and > even after extending the timeout period to something crazy like 6 hours it > still times out. Anyone have any performance suggestions for doing daily > differential calculations? Use calcsize or server-side estimates? see the 'estimate' dumptype parameter. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: 2.6 upgrade file locations
On Mon, Aug 11, 2008 at 9:26 AM, stan <[EMAIL PROTECTED]> wrote:
> Since I install each application in /opt/{APP_NAME}, I'd prefer not to have
> the 2n'r "amanda" levels. In the past I used dirctives like:
>
> --sbindir=/opt/amanda/sbin
You'll want to look at --with-amlibdir, --with-amlibexecdir, and
--with-amperldir.
Dustin
--
Storage Software Engineer
http://www.zmanda.com
Re: Backups to disk and tape
On Sat, Aug 9, 2008 at 9:22 AM, Peter Spikings <[EMAIL PROTECTED]> wrote: > Is that possible? (because I can't see how from reading the man pages). What you're asking about is called D2D2T, and Amanda doesn't support it directly just yet, although preliminary support is a goal for the next release or so. RAIT is the current best option. As long as you're mirroring, there's no problem with deleting a vtape after 7 days. If you need to recover that data, you'll just retrieve the tape and use it directly. So a little bit of shell wrapper can get you up and running with this scheme in no time. I'd be interested to hear folks' high-level thoughts on how D2D2T (or, more generally, data migration) configuration should work in Amanda. Should migration be a separate step from dumping (a la amflush)? Should migration be scheduled somehow (and if so, with what sort of parameters), or should Amanda require explicit instructions on which dumps to migrate and when? We're still laying the low-level groundwork for migration, so there's lots of time to pontificate. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: Theoretical question
On Fri, Aug 8, 2008 at 6:38 PM, Steffan Vigano <[EMAIL PROTECTED]> wrote: > because "the LTO tapes only hold 100GB natively, and you can't backup a file > system with dump that's larger then the actual capacity of the tape." .. > A) Was there any truth to his original statement? Nope -- that was a pre-spanning statement. > B) Since the initial setup we've upgraded Amanda several times (currently on > 2.5.1p2) and we now have tape spanning going. Assuming "A" is true, do we > still need to partition our disk to 100GB slices? or can we have 1 or 2 > huge slices? No. However, note that Amanda's balancing act is much more effective when it has multiple DLEs of approximately the same size to shuffle around, so it probably makes sense to continue using multiple partitions instead of lumping everything into one DLE. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: backup more than 1 time a day
On Fri, Aug 8, 2008 at 2:40 PM, Mister Olli <[EMAIL PROTECTED]> wrote: > is there some way to instruct amanda to backup the maildirs for example > every 2 hours??? Sure, make sure 'usetimestamps' is true, and then run Amanda every 2 hours from your crontab. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: 2.6 upgrade file locations
On Fri, Aug 8, 2008 at 9:11 AM, stan <[EMAIL PROTECTED]> wrote: > But 2.6 seems to be installing things in /opt/amanda/lib/amanda > Tjis is resulting in things like it looking for the config file (which is > installed in /opt/amanda/etc/DailyDump) in /opt/amanda/etc/amanda/DailyDump We did move some things around, mainly because we added a number of libraries and other new files that really didn't belong in $prefix/lib. However, the change to the config dir comes as a surprise to me. Can you verify that amgetconf build.CONFIG_DIR gives /opt/amanda/etc/amanda? If so, can you send along the whole script? Dustin -- Storage Software Engineer http://www.zmanda.com
Re: Busted Tape Drive
This all sounds great. Not to sound like an overworked developer, but could you all please update the FAQ entry accordingly? ;) http://wiki.zmanda.com/index.php/FAQ:Why_can_I_write_new_labels_to_my_tapes_but_can%27t_read_the_old_ones%3F Dustin -- Storage Software Engineer http://www.zmanda.com
Re: Sources for replacement tape drive in Sun L9 array
On Thu, Aug 7, 2008 at 7:37 PM, Craig Dewick <[EMAIL PROTECTED]> wrote: > Can any generic DLT-8000 raw drive be used to replace the existing one? I > believe they're all made by Quantum. Other than the tape drive issue, I've > got Amanda working fine backing up all my systems including my > Strongbolt/CentOS-based Cobalt Raq-4i system (my web server machine). I have[1] an L9, but it shipped with a dead drive. I haven't been able to find a DLT-8000 drive that I *know* will work with the library -- all those pieces-parts to the door and loading/unloading mechanism have to line up just right, and it has to fit into the library's mounting brackets. So I don't have the answer, but I'd be very interested if someone else does! Dustin [1] Well, had. I gave it to Dan Locks. Hi, Dan. -- Storage Software Engineer http://www.zmanda.com
Re: Glibc versions
On Thu, Aug 7, 2008 at 5:27 PM, stan <[EMAIL PROTECTED]> wrote: > I am trying to build 2.6.0p1 on a Ubuntu machine. As far as I know this > machine is the latest version of Ubuntu. The configure script fails on the > check for glibc version 2.2.0 or newer. I am curious if this version of > glibc is so new that it would not be in the Ubuntu stable release yet? Nope -- I think it was released in 1998. They're up around 2.2.16 now. Look at the end of the logging section of config.log -- often this is caused by path problems, or by a missing pkg-config. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: Busted Tape Drive
On Thu, Aug 7, 2008 at 5:25 PM, Steven Backus <[EMAIL PROTECTED]> wrote: > FAQ:Why can I write new labels to my tapes but can't read the old ones? > > Check your blocksize. This behavior occurs when the tape unit is > set to have a particular block size (that may not match at read > time what was used to write a tape). Looks good: http://wiki.zmanda.com/index.php/FAQ:Why_can_I_write_new_labels_to_my_tapes_but_can%27t_read_the_old_ones%3F Are there more details to add? The 'mt' invocations in rc.conf may be helpful to others, for example. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: Busted Tape Drive
On Thu, Aug 7, 2008 at 1:06 PM, Steven Backus <[EMAIL PROTECTED]> wrote: > Thanks to John Hein and everyone. I haven't followed this too closely, but are there some takeaway messages that could be worked into a FAQ here? http://wiki.zmanda.com/index.php/FAQ Dustin -- Storage Software Engineer http://www.zmanda.com
Re: problem with amflush?
The "DUMPS DID NOT COMPLETE" message comes when no FINISH log message (except planner) is seen. If you compare the logfile in the previous emails to your own, successful logfiles, you'll see that line is missing. The most likely cause is a segfault or other bug in the driver. The task, then, is to find it. Andre, can you send along your debug logs for the driver? Also, what version of Amanda are you running on the server? (and as always, my apologies if you've already specified) Dustin -- Storage Software Engineer http://www.zmanda.com
Re: Port xxx is not secure
This funny port number is coming from the NAT operations in your firewall. There are ways to maneuver Amanda around that, but you might be better served to simply use ssh authentication, if that's an option. Otherwise, you should either set up a static NAT on the firewall or set the Amanda client's port ranges "wide open" (1024-65535, or to whatever your firewall's range of dynamically assigned ports happens to be). Dustin -- Storage Software Engineer http://www.zmanda.com
Re: Can't open tape device /dev/nst0
On Wed, Aug 6, 2008 at 7:32 PM, Vectorz Sigma <[EMAIL PROTECTED]> wrote: > I downgraded to Amanda 2.5.2p1 and it gave me a better clue when running > amlabel. Basically it said /dev/nst0 = permission denied. I've been > running the amlabel command as amandabackup but didn't realize you need to > give permissions to the /dev/nst0 device (which was 750 perms and owned by > root:root). I chmod'd to 777 to test and that fixed the problem. > > Is it a known deal that we're supposed to change permission on the tape dev > name? The device should have the group 'disk' (or equivalent), and amandabackup should be in that group. However, it's definitely a bug to not report that as "permission denied", so I'll take a look. Thanks for chasing this down. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: Can't open tape device /dev/nst0
The mystery deepens. Can you run 'amlabel' under strace or truss or your local equivalent? I'd like to know what syscall is returning EBADFD (Bad file descriptor). Also, please remind me what version of Amanda you're running. You can send the strace output to me privately for analysis if you'd like, rather than spamming the list (it will be *large*) Dustin
Re: Can't open tape device /dev/nst0
On Wed, Aug 6, 2008 at 11:00 AM, Vectorz Sigma <[EMAIL PROTECTED]> wrote: > Doyle, thanks for the tip but I got errors when trying it without the > "tape:" in the string. I forgot to mention that I'm using 2.6.0p1 The tape: thing is not the issue here -- the code that writes the deprecation warning adds the prefix, and things go merriliy on. So the error is, in particular: > Can't open tape device /dev/nst0: Bad file descriptor > amlabel: Could not open device /dev/nst0. Which basically amounts to: "the kernel said there was an error." So the next step is to ask the kernel -- check in dmesg, access the device using dd or mt, and so on. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: Can't open tape device /dev/nst0
On Tue, Aug 5, 2008 at 3:04 PM, Vectorz Sigma <[EMAIL PROTECTED]> wrote: > I have an Ultrium LTO-4, and cannot get it to work with Amanda. You had pasted some other diagnostic output that I didn't recognize to #amanda. Can you give some more details? Also, please have a look in your kernel logs to see if any HW errors have been detected. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: Testing for Amanda-2.6.0p2
On Tue, Aug 5, 2008 at 3:45 PM, Gene Heskett <[EMAIL PROTECTED]> wrote: > When was that, Dustin? This has been a minor niggle to me since back in the > 2.5.x days. Am I the only one that actually reads these email reports amanda > can send? You're probably reading them more closely than others :) The bug was introduced in March, which means it was in 2.6.0p1. I didn't look too closely at the code before that, so this bug may also have been present in earlier versions. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: Testing for Amanda-2.6.0p2
On Fri, Aug 1, 2008 at 2:50 PM, Gene Heskett <[EMAIL PROTECTED]> wrote: > Running it every night Dustin & the only wart is the reported compression in > the emailed report, that is generally completely bogus. I have posted about > it, but no one replied AFAIK. OK, my apologies for not getting back to this earlier. This is definitely a bug, introduced in another fix to amreport. I think I have a fix to it, but I want to run it by Jean-Louis before committing it. Whatever happens, we'll fix it for 2.6.0p2. Anyone else have 2.6.0 bugs to report? Dustin -- Storage Software Engineer http://www.zmanda.com
Re: Busted Tape Drive
On Tue, Aug 5, 2008 at 11:09 AM, Steven Backus <[EMAIL PROTECTED]> wrote: > This could be it, does amanda use this for tape commands? Amanda uses what was supplied in the TAPEDEV parameter directly. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: Testing for Amanda-2.6.0p2
On Fri, Aug 1, 2008 at 7:05 PM, Gene Heskett <[EMAIL PROTECTED]> wrote: >>Can you send the trace logfile that generated this report? It should be titled log.200807* and should be in your logdir. Logdir is available from 'amgetconf DailySet1 logdir'. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: Testing for Amanda-2.6.0p2
On Fri, Aug 1, 2008 at 2:50 PM, Gene Heskett <[EMAIL PROTECTED]> wrote: > Typically off just enough to be obvious, occasionally wildly, glaringly > obvious. Most reports aren't this wild though: > > Avg Compressed Size (%) 467.0--52.0 (level:#disks ...) > > Other than that, I don't have any indications of trouble. If I did, you would > already know about it. :) Great -- thanks! Can you send the trace logfile that generated this report? Dustin -- Storage Software Engineer http://www.zmanda.com
Testing for Amanda-2.6.0p2
Amanda Hackers: We're gearing up for a new patch release of 2.6.0 in about 3 weeks, and I'd like to make sure that we get a lot of testing before that happens. I've already sent out an email to the platform experts, and have heard back from a few. We quashed a few bugs this way, and there are a few still outstanding (the SCO OpenServer problem comes to mind). However, I don't have a good feeling for how many folks are using the 2.6.0 release, and what sorts of bugs they've encountered. We've made a lot of fixes to the Device API since 2.6.0p1, all of which need testing. What I need from *you* (yes, I'm looking at you, not the person beside you) is to download and test out the latest 2.6.0 images from http://www.zmanda.com/community-builds.php. You can test these non-invasively with the instructions at http://wiki.zmanda.com/index.php/Testing, or incorporate them into your production systems. Positive or negative, please let me know what you find out. I'm completely available to help, too -- here, on #amanda, or wherever else you need me. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: SELinux attrs
On Tue, Jul 29, 2008 at 1:05 PM, Albrecht Dreß <[EMAIL PROTECTED]> wrote: > Finally, rebuild amanda with the configure option > "--with-gnutar=/usr/sbin/amgtar" (this should be easier with 2.6, where an > application can be defined in the runtime config? Is that correct?). In a > first quick test, this /seems/ to work. I didn't test it thoroughly, so > *please* be careful if you want to go ahead with it on a production system! Yes, easier in 2.6.0. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: SELinux attrs
On Tue, Jul 29, 2008 at 10:22 AM, Nick Smith <[EMAIL PROTECTED]> wrote: > I've just experimented with 2.6.0p1 and found the the Application API *does* > work - cool!! Maybe it's just the configuration for the application and > properties that wasn't included. I was wrong, sorry. It's an early version, and misses a lot of functionality that turns out to be necessary. If it works for you, great! If you encounter bugs or need more features, please consider building against HEAD. The Application API is designed with your needs in mind. We would absolutely *love* to have any feedback you can provide. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: Checking for GLIB error
On Tue, Jul 29, 2008 at 10:15 AM, Nick Smith <[EMAIL PROTECTED]> wrote: > Just a question - how do I report Amanda bugs these days? I've uncovered two > bugs which corrupt memory but only cause crashes with the SunStudio version! Just report them here, or on amanda-hackers if you'd rather limit your audience. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: SELinux attrs
On Tue, Jul 29, 2008 at 6:21 AM, Nick Smith <[EMAIL PROTECTED]> wrote: > Is the Application API in Amanda 2.6.0p1? I'm asking because I cannot > configure my application for ZFS dumping as per description from Zmanda > site? I get the following error when I try to test the configuration: No -- we cut 2.6.0 to *include* the Device API but *exclude* the Application API, on the grounds that adding two significant new chunks of code in one release would introduce too many bugs. The App API is present in trunk, however, and will be in 2.6.1. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: Checking for GLIB error
On Tue, Jul 29, 2008 at 5:12 AM, Nick Smith <[EMAIL PROTECTED]> wrote: > FYI I was able to get Amanda 2.6.0p1 to compile on OpenSolaris 2008.05 > *without* any external dependencies! I used the SunStudio 12 and the > GLIB 2.0 that's supplied with GNOME Desktop Developer packages (so there > still some dependencies). Great! This reflects my understanding that there are two ways to build 2.6.0 on Solaris: with gcc + a gcc-based glib, or with SunStudio + the built-in glib. Would you mind updating the relevant wiki page to make this clearer to future users? http://wiki.zmanda.com/index.php/Installation/OS_Specific_Notes/Installing_Amanda_on_Solaris Dustin -- Storage Software Engineer http://www.zmanda.com
Re: SELinux attrs
On Mon, Jul 28, 2008 at 12:52 PM, Albrecht Dreß <[EMAIL PROTECTED]> wrote: > A while ago I started to write a simple wrapper script for star (you have to > re-build amanda, as the "tar" application it calls is hard-coded in the > executable), but did not yet succeed. I probably have to look deeper into > the sources to see which options might be used with gtar, so the proper star > replacements are mapped. Of course, it would also make sense to include a > "star" backup mode into amanda, as the extended attributes problem ist > rather common. If anyone else did that work already, I would highly > appreciate any pointers! You should take a look at the application API -- it will make such a creation much easier! Dustin -- Storage Software Engineer http://www.zmanda.com
Re: directory locations and dumpuser after new install of amanda 2.6.0p1
On Sat, Jul 26, 2008 at 11:08 AM, Doyle Collings <[EMAIL PROTECTED]> wrote: > When you say thay your getting rid of --with-maxtapeblocksize in > Amanda-2.6.1, does that mean I will be able to specify blocksize in my > tapetype definition without it? Is is neccesary now or does it just specify > the maximum blocksize I can use? Is there a source RPM for 64bit SLES? a) yes b) yes, it currently just specifies the maximum size (well, it's kinda complicated, but) c) Source RPMs are architecture-agnostic. To my knowledge, our RPMs will build on SLES10 - x86_64. If you have trouble, post to the list. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: directory locations and dumpuser after new install of amanda 2.6.0p1
On Sat, Jul 26, 2008 at 2:43 AM, Doyle Collings <[EMAIL PROTECTED]> wrote: > I have decided to clean out my old installations of Amanda 2.46, 2.6.0p1 rpm, > and 2.6.0p1 from source and start fresh. I noticed that the defaults of the > install from source do not match the rpm from zmanda. For example, The link > http://wiki.zmanda.com/index.php/Amanda_packages_from_Zmanda_downloads_page > states that the Amanda commands are in /usr/sbin. Installing from source > puts them in /usr/local/sbin. It appears that installing from source still > puts the amanda user as the default dumpuser and not amandabackup. Bottom > line, I want my install from source directories and default user to matchup > with intended direction of present and future packages. I would use the RPM > provided but I want to configure my installation configure > --with-maxtapeblocksize=N. I also am running the 64 bit version of SLES > 10SP1 and want to compile with the 64 bit code. Your best bet may be to take the source RPM and tweak the .spec file tp specify your --with-maxtapeblocksize, then build your own RPM. FWIW, we're getting rid of --with-maxtapeblocksize in Amanda-2.6.1. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: howto fix over-write email information
On Thu, Jul 24, 2008 at 3:56 AM, Snorre Stamnes <[EMAIL PROTECTED]> wrote: > Is there a way to fix this? See the "columnspec" config parameter >> (brought to you by Amanda version 2.4.5p1) I don't know if the parameter was present back then. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: Is there a log that tells me the block size of a Backup Job
On Thu, Jul 24, 2008 at 4:44 PM, Doyle Collings <[EMAIL PROTECTED]> wrote: > What do I need to do next to troubleshoot a misconfiguration or hardware > problem. If I am using a bad installation, I am not averse to wiping the > machine clean of Amanda and re-installing from source.. Try setting up a more typical Amanda configuration -- a DLE with more than one file in it, multiple DLEs, a holding disk, and so on -- and see if you can get that working. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: Checking for GLIB error
On Thu, Jul 24, 2008 at 11:22 AM, McGraw, Robert P <[EMAIL PROTECTED]> wrote: > I assume that this is not working because of the compile error. I assume > that the invalid option is the -mt. > > I am not sure what is causing this option in be inserted. Any suggestions. http://wiki.zmanda.com/index.php/Installation/OS_Specific_Notes/Installing_Amanda_on_Solaris#GNU_gcc In particular, you need to use the same compiler that glib was compiled with. The reason is that glib supplies compiler flags via pkg-config. That's where the -mt option is coming from. Blastwave compiles with the Sun compiler, so you'll need to use that. Compiling on Solaris is annoying, but I'm probably preaching to the choir on that point :) Dustin -- Storage Software Engineer http://www.zmanda.com
Re: Is there a log that tells me the block size of a Backup Job
On Thu, Jul 24, 2008 at 2:01 AM, Doyle Collings <[EMAIL PROTECTED]> wrote: > I am only backing up one file, an iso image, which is about 2.6 gigabytes in > size. Maybe that is why I get an error after 2.5 gigs. The tapes The LTO4 > tape drive I am using is has a capacity of 800 gigabytes uncompressed and 1.6 > terabytes compressed. I have tarred to that drive at speeds of 5.5 gigabtyes > a minute. I don't have a holding disk configured because my experience with > tar led me to believe that my tape could handle the speed of my fastest > source drive. I figured the holding disk would only slow me down. I have > corresponded with others that use LTO3 drives that are having faster speed > than me. I don't know if any of them are using gnu-tar. One who has been > using Amanda since the early Ninetees is definitley still using dump. Is my > holding disk logic flawed? I am only backing up from one local fibre > attached SAN Well, a few things, in order: 1. You have some kind of misconfiguration or hardware problems if your taper is erroring out before the tape is full. First priority is to fix that. 2. Backing up a single file isn't really a good use of Amanda, and *definitely* is a waste of 797.4GB of tape space. I assume this is a "test" arrangement for Amanda, and that you'll be backing up more data later? If so, you will get more realistic results from a more realistic test. 3. Your measurements comparing Amanda to tar are flawed. If you can successfully write a larger dump file (tens of GB) to tape, we can look at the tape write rate and determine whether and by how much Amanda's tape writes are slower than tar. 4. You can definitely optimize Amanda for a better tape write speed. Adding a holding disk would be one way to do that. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: Is there a log that tells me the block size of a Backup Job
On Wed, Jul 23, 2008 at 9:53 PM, Ian Turner <[EMAIL PROTECTED]> wrote: > This appears to be a bug in Amanda. I can't reproduce it on my end, though; is > there any chance you can run amcheck in a debugger with these options and see > what you get? If you break on get_int and generate a backtrace from there, it > may provide some clues as to what is going on. The configuration bug got solved -- it was a missed merge from trunk to the 2.6.0 branch. At this point, the question is what's going on with the tape drive, and I think that has more to do with Doyle's local configuration, although I'll never quite rule out a bug in Amanda. :) Dustin -- Storage Software Engineer http://www.zmanda.com
Re: Is there a log that tells me the block size of a Backup Job
It looks like you don't have a holding disk configured, so your writes to tape are limited to the speed of the dump. You should configure a holding disk. Also, the tape drive gave an error after only ~2.5G. I'm not sure why that would happen -- your tapes are 80G, right? The taper log may have some additional information in it. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: amanda-2.5.2p1 on SCO 5.0.4
On Wed, Jul 23, 2008 at 2:38 PM, Michael Reuland <[EMAIL PROTECTED]> wrote: > I have stripped the ~etc/amanda/amanda-client.conf file to only the actual > configuration keywords and options. I still get the similar response > reporting the first line of configuration data from the file. What happens if you switch the order of lines so that the first line is e.g., the second? > Is there a utility that I can run on the client to see if the config file is > being read correctly? I'd like to narrow down the problem. Since I have > the host server running w/ a different client (not SCO), I'd like to focus > on the client platform to start with. You can run 'amrecover' on the client. It should give the same error messages. On a completely unrelated topic, I'd like to know whether your SCO system uses POSIX tapes or the old "uware" driver. Can you run grep 'tape-.*\.c' device-src/Makefile and send me the result? I'm considering removing native uware (and aix and xenix) support, since most systems have POSIX these days, and since we don't have a way to test these other systems. Dustion -- Storage Software Engineer http://www.zmanda.com
Re: Is there a log that tells me the block size of a Backup Job
On Wed, Jul 23, 2008 at 3:49 PM, Dustin J. Mitchell <[EMAIL PROTECTED]> wrote: > This is all very strange. Did you compile from source? If so, can I > get you to put some dbprintf's in apply_config_overwrites and send me > the results? Better, is there any chance you could hop into #amanda > on IRC so we can figure this out? OK, we worked this out in IRC. The deal with the -o options is that there was a fix that wasn't properly ported to the 2.6.0 branch. The fix will be included in 2.6.0p2. As for the block size and timing issues, Doyle's looking at overall runtime of tar vs. overall runtime of Amanda, which isn't a fair comparison. Amanda is: - making estimates - planning - appying compression - sending the dump over a network connection (on the loopback interface, but still) - creating an index - creating an incremental file - spooling to holding disk - writing a header file to the tape device I expect that the actual taper time will be closer to the duration of the tar invocation. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: sizeof error for 2.6.0p1 compile
On Wed, Jul 23, 2008 at 3:52 PM, McGraw, Robert P <[EMAIL PROTECTED]> wrote: > I tried to compile amanda-2.6.0p1 on a Sparc Solaris 9 & 10 host and get the > following error: The actual error in config.log (sent privately) is configure:49348: /pkgs/gcc-3.4.3/bin/gcc -o conftest -g -O2 -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64 -D_REENTRANT -I/opt/csw/include -D_GNU_SOURCE -I/opt/csw/include -L/opt/csw/lib conftest.c -lcrypto -L/opt/csw/lib -lcurl -L/opt/csw/lib -L/opt/csw/lib -lidn -lssl -lcrypto -llber -lldap -lsocket -lnsl -lz -lnsl -lsocket >&5 configure:49351: $? = 0 configure:49357: ./conftest ld.so.1: conftest: fatal: libcurl.so.4: open failed: No such file or directory I'm guessing this is related to Solaris' habit of requiring both a runtime and compile-time library search path be specified. You can probably fix this with ./configure LDFLAGS='-R/opt/csw/lib' ... Dustin -- Storage Software Engineer http://www.zmanda.com
Re: Is there a log that tells me the block size of a Backup Job
On Wed, Jul 23, 2008 at 5:02 PM, Doyle Collings <[EMAIL PROTECTED]> wrote: > I tried those with runtapes=4 and 'runtapes=4' with the same results > how do I hop into #amanda on IRC You can actually use a web-based client: http://www.mibbit.com/ The network is freenode.net, and the channel is #amanda. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: Is there a log that tells me the block size of a Backup Job
On Wed, Jul 23, 2008 at 3:40 PM, Doyle Collings <[EMAIL PROTECTED]> wrote: > Yes, same error. What should I put quotes to treat it as an integer? > > [EMAIL PROTECTED]:/root> /usr/local/sbin/amdump fullback -o runtapes=1 > argument "": an integer is expected > > ** (process:9825): CRITICAL **: parse error in configuration overwrites This is all very strange. Did you compile from source? If so, can I get you to put some dbprintf's in apply_config_overwrites and send me the results? Better, is there any chance you could hop into #amanda on IRC so we can figure this out? Other things to try: /usr/local/sbin/amgetconf Conf -o runtapes=4 runtapes /usr/local/sbin/amgetconf Conf -o 'runtapes=4' runtapes Dustin -- Storage Software Engineer http://www.zmanda.com
Re: sizeof error for 2.6.0p1 compile
On Wed, Jul 23, 2008 at 3:52 PM, McGraw, Robert P <[EMAIL PROTECTED]> wrote: > checking size of int... configure: error: cannot compute sizeof (int) > See `config.log' for more details. > > Any suggestions on what is causing this? The error is a bit misleading -- this just happens to be the first thing that the configure script tries to compile, so if it fails for unrelated reasons, you'll get this message. config.log consists of a long log of configure's activities, followed by some long dumps of various variables and whatnot. Look at the end of the activity log to find the compiler error that caused the compile to fail. If it's not obvious how to fix it, post it here and we'll take a look. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: Is there a log that tells me the block size of a Backup Job
On Wed, Jul 23, 2008 at 11:00 AM, Doyle Collings <[EMAIL PROTECTED]> wrote: > I wasn't using that extra space between = and 24. Does the "** > (process:17472): CRITICAL **: parse error in configuration overwrites" give > us any clues. Would it help if I cleaned out the installation and recompiled? > Two earlier rpm versions of amanda were installed previously. Well, it just says that your configuration overwrite (-o option) is badly formed. Maybe your shell is doing funny things -- try quoting it? Also, try a few things such as amdump -o runtapes=1 to see if they give a similar error Dustin -- Storage Software Engineer http://www.zmanda.com
Re: usage of holding disk
On Wed, Jul 23, 2008 at 9:16 AM, Andre Brandt <[EMAIL PROTECTED]> wrote: > I've configured our tape backup with amanda a few weeks ago. Everything is > fine - the backup is done every night without any problems. The problem is, > that our office is closed from time to time. And if the tape isn't changed, > amanda waits until the right tape is inserted and after that the backup > starts. Is it possible to configure amanda to anyway start the backup and > use the holding disk although right the tape isn't inserted? So that, for > example, one or two backup jobs are written to disk and after the tape were > inserted, they will be written to tape? I assume you're using chg-manual? There are some timeout knobs you can turn in that script, which are set to very long durations (days) by default. If you dial those down to, say, a few hours, and make sure that when you *are* present, you change the tapes within that time, you should be OK. By the way, even while Amanda is waiting for a tape, it is spooling dumps to holding disk. If you know the days the office will be unstaffed, you can also run amdump Conf -o tapedev= -o tpchanger= to avoid even requesting a tape. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: Is there a log that tells me the block size of a Backup Job
On Wed, Jul 23, 2008 at 1:30 AM, Doyle Collings <[EMAIL PROTECTED]> wrote: > [EMAIL PROTECTED]:/> /usr/local/sbin/amdump fullback -o > tapetype:LTO4:blocksize=2048 > > results: > argument "": an integer is expected > argument "": Tape blocksize must be at least 32 KBytes > ** (process:17472): CRITICAL **: parse error in configuration overwrites The only way I can reproduce that error is to write (note the extra space): [EMAIL PROTECTED] ~/devel/projects/amanda/t/amanda $ amdump Conf -o tapetype:DISK:blocksize= 2048 argument "": an integer is expected argument "": Tape blocksize must be at least 32 KBytes errors processing config file at /A/p/sbin/amgetconf line 275. -- Storage Software Engineer http://www.zmanda.com
Re: Is there a log that tells me the block size of a Backup Job
On Tue, Jul 22, 2008 at 6:51 PM, Doyle Collings <[EMAIL PROTECTED]> wrote: > Is there a way to manually define the tapetype from the command line and echo > the results of each command? amdump -o tapetype=mytapetype or maybe you mean amdump -o tapetype:LTO4:blocksize=2048 > Is there some kind of config command to check a configuration status? Not sure what you mean, beyond what amcheck does. If you want to see, for absolutely sure, what blocksize it's using, make up a fairly small backup set (trim your disklist temporarily) and run amdump under 'strace -f -e write amdump 2>/tmp/trace' or the equivalent for your operating system. You should see a bunch of write commands in /tmp/trace, and the size of the writes are the blocksize Amanda is using. There could be lots of other things slowing Amanda down. Are you doing software compression? Are you using a holding disk? Dustin -- Storage Software Engineer http://www.zmanda.com
Re: amanda on Mac OS X (10.5.4) Leopard
On Tue, Jul 22, 2008 at 3:11 PM, myron <[EMAIL PROTECTED]> wrote: > This may take a while. I went to the step on setting up the amanda user and > the niutil command is no longer there. Read the whole page - there's a fix for that already. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: amanda on Mac OS X (10.5.4) Leopard
On Tue, Jul 22, 2008 at 1:42 PM, myron <[EMAIL PROTECTED]> wrote: > I looked at > http://wiki.zmanda.com/index.php/Installation/OS_Specific_Notes/Installing_Amanda_on_Mac_OS_X#Version-Specific_Notes > and it doesn't mention leopard. Do I follow the same instructions? As far as I know. My machines are still 10.4. I'm cheap :) > I noticed that on my Mac /etc./services already has amanda in it. > amanda 10080/udp # Amanda > amanda 10080/tcp # Amanda > # John Jackson <[EMAIL PROTECTED]> > # <[EMAIL PROTECTED]> > > The ports listed are different than the wiki instructions. Is there another > page that I should be looking at? And so are the names. The ports in the wiki are only useful if you're using kerberos. You don't need to add those. Please add any notes, or just a description of your success, to the wiki when you get things working. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: amanda-2.5.2p1 on SCO 5.0.4
On Mon, Jul 21, 2008 at 12:24 PM, Michael Reuland <[EMAIL PROTECTED]> wrote: > WARNING: apd50.coop.com: selfcheck request failed: tcpm_recv_token: invalid > size: "/usr/local/etc/amanda/amanda-client.conf", line 5: a quoted string is > expected > "/usr/local/etc/aman > Client check: 2 hosts checked in 0.161 seconds, 1 problem found This is basically a protocol error -- the server connected to the client, expecting a binary-formatted packet in response, but the config parser output an error message over the socket instead. > The client system apd50.coop.com (SCO 5.0.4) shows the following when I try > to run amcheck on that system. Running amcheck on a client doesn't make sense, and won't parse the amanda-client.conf. > The ~etc/amanda/amanda-client.conf file is as follows: > > # amanda.conf - sample Amanda client configuration file. > # > # This file normally goes in /etc/amanda/amanda-client.conf. > # > conf "DailySet1" # your config name > > #index_server "myserver.mydomain.net" # your amindexd server > index_server "myserver" # your amindexd server > > #tape_server "myserver.mydomain.net" # your amidxtaped server > tape_server "myserver" # your amidxtaped server > > #tapedev "/dev/null" # your tape device > # auth - authentication scheme to use between server and client. > # Valid values are "bsd", "bsdudp", "bsdtcp" and "ssh". > # Default: [auth "bsdtcp"] > > auth "ssh" > > # your ssh keys file if you use ssh auth > > ssh_keys "/usr/amanda/.ssh/id_rsa_amrecover" Unless I'm missing something, line 5 is the "conf" line. Try commenting that line out? > The configure line that I used for the SCO build was: > > configure --program-prefix= --bindir=/usr/local/bin \ > --includedir=/usr/include --libdir=/usr/lib --libexecdir=/usr/lib/amanda \ > --localstatedir=/var/lib --with-gnutar-listdir=/usr/local/bin/gtar \ > --with-user=amandabk --with-group=backup --with-tmpdir=/var/log/amanda \ > --enable-shared --enable-static --enable-threads --without-ipv6 > > I've check all the files and directories and confirmed that all they are all > owned and writable by amandabk w/ the few exceptions of the files that have > SUID root. Please post suggestions to help others build on SCO on the wiki: http://wiki.zmanda.com/index.php/Installation/OS_Specific_Notes and also please consider becoming a Platform Expert: http://wiki.zmanda.com/index.php/Platform_Experts We don't have anyone else around to test on SCO, so your help is the only way we can continue to compile on this system. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: Amanda 2.6 on OpenSolaris 2008.05?
On Sat, Jul 19, 2008 at 1:34 PM, Paul Crittenden <[EMAIL PROTECTED]> wrote: > I have installed on Solaris 9 and just yesterday on Solaris 10. I have some > environment variable settings and configure settings I used. However, they > are at work and I am at home. If you would like or need them I can probably > remote in and get them for you. Was this 2.6.0p1? I'd be interested to hear about any issues you encountered, as we're gearing the 2.6.0 branch up for another patch release. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: Proble with amanda and MacoS X
Good to hear you're building on a Mac! Have you set up the necessary launchd script on the mac? There's information about it at http://wiki.zmanda.com/index.php/Installation/OS_Specific_Notes/Installing_Amanda_on_Mac_OS_X Dustin -- Storage Software Engineer http://www.zmanda.com
Re: [Amanda-users] Full/Incremental and retension
On Fri, Jul 18, 2008 at 6:48 PM, Debux <[EMAIL PROTECTED]> wrote: > correct? Yes. > And how does the retention? In the next cycle (next week) will be overridden > in the data used tapes or recording sequence occurs? Amanda's retention policy is based exclusively on the tape cycle. Data is retained until the tape it is on is overwritten, at which point it's gone. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: [Amanda-users] Full/Incremental and retension
On Fri, Jul 18, 2008 at 2:43 PM, Debux <[EMAIL PROTECTED]> wrote: > I want to help set up the Amanda backup for weekly full, incremental, in the > other day and retensão of two weeks. > I have 6 tapes LTO3. > > This is possible? > > My setup so far is this: > > dumpcycle 1 weeks > runspercycle 7 days > tapecycle 6 tapes > runtapes 1 Your runspercycle has to be less than your tapecycle, so you'll either need more tapes or fewer runs. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: amdump return code quesion
On Fri, Jul 18, 2008 at 9:05 AM, McGraw, Robert P <[EMAIL PROTECTED]> wrote: > I did an exit code check of my amdump last night and I got an exit code 8 > "Don't know the status of a dle (RESULTS_MISSION in the report)" > > Can some one tell me what this means. I think you meant "RESULTS_MISSING" :) It means that at least on DLE failed to back up. You should check the report for more details. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: how to check for a good amanda run
On Thu, Jul 17, 2008 at 12:55 PM, McGraw, Robert P <[EMAIL PROTECTED]> wrote: > It seem that the local man page did not have anything about exit codes for > amdump, but the zmanda man pages did. I believe I got what I needed. We just added that to the manpages a while ago, so no worries. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: how to check for a good amanda run
On Thu, Jul 17, 2008 at 10:35 AM, McGraw, Robert P <[EMAIL PROTECTED]> wrote: > How can I determine the status of an amdump run? Amdump has a rather detailed exit status. See http://wiki.zmanda.com/man/amdump.8.html Dustin -- Storage Software Engineer http://www.zmanda.com
Re: another selfcheck problem
The debug logs *do* look different, so this was at least a minimal improvement: On Wed, Jul 16, 2008 at 2:26 PM, Mister Olli <[EMAIL PROTECTED]> wrote: > amandad: debug 1 pid 80108 ruid 1002 euid 1002: start at Wed Jul 16 > 20:21:46 2008 vs. (earlier) > -auth=bsd: debug 1 pid 76969 ruid 1002 euid 1002: start at Wed Jul 16 10:33:47 2008 What do the selfcheck logs say? Dustin -- Storage Software Engineer http://www.zmanda.com
Re: another selfcheck problem
On Wed, Jul 16, 2008 at 9:46 AM, Mister Olli <[EMAIL PROTECTED]> wrote: > amanda dgram udp waitamanda /usr/local/libexec/amandad > -auth=bsd amdump amindexd amidxtaped It looks like amandad is thinking that -auth=bsd is its name, which means that it saw that in argv[0] Try amanda dgram udp waitamanda /usr/local/libexec/amandad /usr/local/libexec/amandad -auth=bsd amdump amindexd amidxtaped Dustin -- Storage Software Engineer http://www.zmanda.com
Re: How Do I Back Up a Smartstor?
Thanks, Jon -- somehow I read "back up TO this beast", which is an entirely different issue :) Dustin -- Storage Software Engineer http://www.zmanda.com
Re: tar vs. dump for / backup?
On Tue, Jul 15, 2008 at 2:23 PM, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote: > What do I do about it? > What does it mean? > Is this a manifestation of the reiserfs-dump incompatibility? Perhaps -- but you said /boot was reiserfs too, and was backed up fine? Are you sure both partitions are reiserfs? Dustin -- Storage Software Engineer http://www.zmanda.com
Re: How Do I Back Up a Smartstor?
On Tue, Jul 15, 2008 at 1:15 PM, Steven Backus <[EMAIL PROTECTED]> wrote: > Description: SMARTSTOR NS4300N 4BAY NAS SATA RAID5 SOHO/SMB .. > And am wondering how one would back up this beast with amanda? I would think you would format it into vtapes, assuming it supports NFS or something else "mountable". Dustin -- Storage Software Engineer http://www.zmanda.com
Re: tar vs. dump for / backup?
On Tue, Jul 15, 2008 at 1:11 PM, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote: > Searching the web for 'amanda all estimate failed' has turned up many > hits, but few fixes. Is this a broad error, or could some fix for it go > in the FAQ? It's a broad error: essentially "something funny happend on the client, while I was asking for estimates" Check the sendsize debug log files on the client in question. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: Why are there log files in /etc/amanda/DailySet1?
On Fri, Jul 11, 2008 at 10:33 AM, Zanga Chimombo <[EMAIL PROTECTED]> wrote: > i should've said that my rant is against the documentation not the design > rationale. ..it *is* on a wiki, nudge nudge. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: debian client
On Thu, Jul 10, 2008 at 11:32 PM, Ian Turner <[EMAIL PROTECTED]> wrote: > If it's not there, you'll can look at the contents of the libamanda.so shared > library using the strings command. That file will name a directory called > AMANDA_TMPDIR; the client should create a subdirectory 'amanda' of that. That > is, if libamanda.so contains the string 'AMANDA_TMPDIR="/foo/bar"', then the > Amanda client will create debugging logs in /foo/bar/amanda. You can also run 'amgetconf build.AMANDA_DBGDIR' Dustin -- Storage Software Engineer http://www.zmanda.com
Re: Why are there log files in /etc/amanda/DailySet1?
On Fri, Jul 11, 2008 at 2:55 AM, Zanga Chimombo <[EMAIL PROTECTED]> wrote: > the following line in amdump.X seems to give a clue: > "dumper: stream_client: connect to 192.168.0.60.46830 failed: No route to > host" > why is the server trying to dump to the client? Since Jon and Paul have addressed the first half of your email so accurately, I guess I'll take this half. Amanda functions by connecting from the server to the client. So this is the dumper connecting to the client to *pull* the data to be backed up. You don't sound surprised about the "No route to host" part, so I assume there's a firewall or other intervening mechanism that would prevent this sort of activity. You'll need to remove this mechanism and allow the server to connecto to the client. Also, given the port number, it sounds like you're using BSD authentication, which tends to use a lot of separate connections instead of one single connection. If your client supports it, consider switching to bsdtcp authentication. See the wiki for a bit more detail. Dustin -- Storage Software Engineer http://www.zmanda.com
Re: Dump & Write
On Wed, Jul 9, 2008 at 5:25 PM, Mike Fahey <[EMAIL PROTECTED]> wrote: > Can I force amanda to write the dumps it has finished while it is still > dumping? Amanda already does that! Note that Amanda can only write one dump to the device (even hard disk) at any given time. Still, that should go pretty quickly. Can you give some more evidence that the taper isn't starting until all of the dumps are finished? Dustin -- Storage Software Engineer http://www.zmanda.com
Re: Amrecover: Header Problem
2008/7/3 Weiwei Luo <[EMAIL PROTECTED]>: > amrecover: Expected: "lev" Actual: "" > WEIRD file > amrecover: bad header > Extractor child exited with status 1 This is the result of an obscure re-entrancy bug that's been fixed in the latest releases. Can you try 2.6.0p1 on that machine? Dustin -- Storage Software Engineer http://www.zmanda.com
Incremental Backups
Im trying to setup incremental backups. I would like to perform 1 full backup during a 7 day period and then incremental the following days. Here is my config. dumpcycle 7 days runspercycle 1 tapecycle 15 tapes Here is my disklist amoverview DailySet1 date 07 host disk 07 65.209.1 /etc 11 65.209.1 /home00 65.209.1 /mysql-backups 11 65.209.1 /var/www 11 any help would be great.
vtape sizes
Im trying to backup to some virtual tapes. My hard drive is 1.2 tb. The /home directory on the remote host is 4 gigs. I want to run the backup every night and do a full backup on the first night. I want to have a 7 day rotation. When I run the backup I get this error. 65.0.0.0 /home lev 0 FAILED "[dump larger than available tape space, 2088825 KB, but cannot incremental dump new disk]" This is what I have. dumpcycle 7 days runspercycle 7 tapecycle 7 tapes
amrecover on ubuntu
Im using amrecover on a ubuntu server. amdump and amcheck work great. When I try to run amrecover I get this error. AMRECOVER Version 2.5.1p1. Contacting server on localhost ... [ ][request failed: timeout waiting for ACK]
Re: Black nasty sex pictures
On Thu, Jul 3, 2008 at 9:32 PM, Gene Heskett <[EMAIL PROTECTED]> wrote: > Dustin, shut this off before I start feeding it to sa-learn -spam. Apologies for a potential duplicate here, but Todd Kover, Jon LaBadie, and the Zmanda IT folks are already on the case. Dustin -- Storage Software Engineer http://www.zmanda.com
