Hi, 13.09.2007 22:14,, root wrote:: > On Thu, Sep 13, 2007 at 09:44:25PM +0200, Arno Lehmann wrote: >> Hello, >> > Thank you for taking the time to reply. > >> 13.09.2007 20:59,, Mike Mestnik wrote:: >>> Am i crazy or should "mount jobid=16" cause that job to >>> unmount the just mounted volume and complain that I need to label. >> Which version of Bacula? >> > *version > mgmt8-dir Version: 1.38.5 (18 January 2006) > > > Connecting to Storage daemon vault7-sd at mgmt8.mn2.visi.com:9103 > > vault7-sd Version: 2.0.3 (06 March 2007) i686-pc-linux-gnu redhat (Stentz) > Daemon started 06-Sep-07 13:49, 3 Jobs run since started. > Heap: bytes=213,859 max_bytes=280,772 bufs=87 max_bufs=98
Oh oh... 1.38 and 2.0 are a bit too different, I suppose. I suggest to upgrade your whole setup to 2.2.3, as this will fix some bugs that might (or even will) get you into trouble sooner or later. >> Can you produce some console output showing the state of the DIR and >> SD before, while and after you do the mount? >> > Here we go... > > While is not possible, the mount only lasts so long. This is a loop > for me, so before and after should match, but I'll print both. > ... ok, this looks more or less normal to me, i.e. I don't see anything explaining the issues you have. >>> Any method of mounting causes the jobid to sqwak about not >>> having a volume, and the device is unmounted. >> So there's no volume available. >> > There have been many volumes, the current is test003. > *llist volume > Pool: Vault7 > MediaId: 1 > VolumeName: test003 > Slot: 0 > PoolId: 1 > MediaType: vault7-File > FirstWritten: 0 > LastWritten: 0 > LabelDate: 2007-09-13 10:25:39 > VolJobs: 0 > VolFiles: 0 > VolBlocks: 0 > VolMounts: 0 > VolBytes: 1 > VolErrors: 0 > VolWrites: 0 > VolCapacityBytes: 0 > VolStatus: Append > Recycle: 1 > VolRetention: 31536000 > VolUseDuration: 0 > MaxVolJobs: 0 > MaxVolFiles: 0 > MaxVolBytes: 0 > InChanger: 1 > EndFile: 0 > EndBlock: 0 > VolParts: 0 > LabelType: 0 > StorageId: 1 > > Pool: VaultX > No results to list. > >>> When I use the File backend how do I select what volume to mount? >> You don't, usually :-) >> > For testing? > A feature request? Of course, if you can give some reasons why this would be necessary. I don't think this is needed for testing; if you want to test the autochanger handling, you can set up a "fake" autochanger, which is controlled by the usual mtx-changer script interface, but uses files instead of tapes. >>> The best I can do is cancel the job, delete the volume. ruining the >>> job then asks for a label, the the label auto-mounts and the job says >>> to use the label cmd(again!) and unmount reports the volume is >>> unmounted already. >>> Unfortunalty that's the loop I'm in. >>> >>> I'm using these instructions: >>> http://www.bacula.org/dev-manual/Brief_Tutorial.html >>> but I'm altering the config for my setup. >> So what does your setup look like now? >> > This link worked for me: > http://sourceforge.net/mailarchive/attachment.php?list_name=bacula-users&message_id=20070913185934.GK8912%40visi.com&counter=1 > >>> autolabel would be nice, if it weren't for this problem. I turned >>> autolabel and every thing else off to get things working again. >>> >>> I was hoping to find some one who has seen this. >>> >>> The SD is running on Linux and the FD and DIR is on FreeBSD... are >>> there any version mismatches I should look out for? >> The only "officially" supported combination is DIR, SD and FD of the >> same version. That said, most often you can use older FDs (unless you >> rely on new features), but the DIR and SD versions should match. >> > That may be difficult, due to vendor supplied patches. I'd prefer > binary over source, save browsing CVS. > > Availability for different OSs may make this difficult or > impossible, for instance MSwin32 is not something I'd ever > learn to build but is something I need to consider supporting. Then I can only suggest to upgrade to 2.2.3, and if that's really not possible, don't use Bacula. I don't think anyone can realistically support a combination of different versions of DIR and SD. >>> Eventually only the DIR will be on FreeBSD. >>> >>> I'm also looking to do multiple directors, one for each class of >>> customer, and they may need to share SDs I could not get multiple >>> Directors to work in the same SD. >>> >>> I'll attach some config. I don't have access to the SD at this time, >>> I'm using 'ssh -i "key"' inside inetd to run 'nc' on the sd server. >>> If there is an interest I am posting this config here. >> I really don't understand what you say here... >> > The communications pass like so... > Dir -localhost:bacula-sd> inetd(exec ssh -vault7:22> sshd(exec nc > -localhost:bacula-sd> bacula-sd)) > Both bacula processes are unaware that they are not connecting > over a localhost:bacula-sd connection. nc(AKA netcat) is used > to proxy the request to the bacula-sd as though it were the > director, inetd accepts connections from the Director as though > it were the sd. Interesting setup... I assume this is to provide some sort of secure communication, similar to what (earlier, for Bacula) was done using stunnel or could be achieved both by using encrypted communications in Bacula, or by setting up a VPN... >>> bacula-sd stream tcp nowait root /usr/bin/ssh -q -T -o >>> BatchMode=yes -i /etc/ssh/vault7-sd_dsa_key [EMAIL PROTECTED] >>> >>> # note for BatchMode to work the server's key needs to be added to the >>> # "system's" known_hosts file, you can also use DNS(but don't quote >>> # me.) ""you will need BatchMode"" so you need to copy the key. >>> >>> vault7:~visi/.ssh/authorized_keys >>> from=mgmt8,command="nc localhost bacula-sd",<othersecureoptions> <the >>> dsa key> >> If I read this correctly, this won't work. The SD needs to run all the >> time. The inetd-setup will probably invoke a new SD for each connection. >> > You can see this is not a problem and that this is my 3rd attempt... > From the sd: 3 Jobs run since started. Right. Still - good luck with this setup ;-) >> Arno >> >>> ------------------------------------------------------------------------ >>> >>> ------------------------------------------------------------------------- >>> This SF.net email is sponsored by: Microsoft >>> Defy all challenges. Microsoft(R) Visual Studio 2005. >>> http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ >>> >>> >>> ------------------------------------------------------------------------ >>> >>> _______________________________________________ >>> Bacula-users mailing list >>> Bacula-users@lists.sourceforge.net >>> https://lists.sourceforge.net/lists/listinfo/bacula-users >> -- >> Arno Lehmann >> IT-Service Lehmann >> www.its-lehmann.de >> >> ------------------------------------------------------------------------- >> This SF.net email is sponsored by: Microsoft >> Defy all challenges. Microsoft(R) Visual Studio 2005. >> http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ >> _______________________________________________ >> Bacula-users mailing list >> Bacula-users@lists.sourceforge.net >> https://lists.sourceforge.net/lists/listinfo/bacula-users >> -- Arno Lehmann IT-Service Lehmann www.its-lehmann.de ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2005. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
