Re: [Bacula-users] Bacula restore runs forever - bootstrap creation [solved]
Martin Simmons schrieb: Ok, that explains the duplicate bootstrap rows. What is the size of the job in bytes? exactly 28321715406 B (approx 27 GB) Did it complete without errors? Yes, it did. Have you run any tools like bscan with this volume to update the catalog? I did! I still was in the phase of evaluating and I tried a lot of things. Several older sessions where already purged. But I can't say for sure WHEN I tried bscan. What does the following slightly modified query print for that jobid? SELECT VolumeName, MediaType, VolIndex, JobMediaId, FirstIndex, LastIndex, StartFile, JobMedia.EndFile, StartBlock, JobMedia.EndBlock FROMJobMedia, Media WHERE JobMedia.JobId=NNN AND JobMedia.MediaId=Media.MediaId ORDER BY VolIndex,JobMediaId; ++---+--+++---+---+-++--+ | VolumeName | MediaType | VolIndex | JobMediaId | FirstIndex | LastIndex | StartFile | EndFile | StartBlock | EndBlock | ++---+--+++---+---+-++--+ | KYE713L4 | LTO-4 |1 |356 | 1 |101610 | 246 | 246 | 0 |77503 | | KYE713L4 | LTO-4 |2 |357 | 101610 |116651 | 247 | 247 | 0 |77503 | | KYE713L4 | LTO-4 |3 |358 | 116651 |126690 | 248 | 248 | 0 |77503 | | KYE713L4 | LTO-4 |4 |359 | 126690 |151467 | 249 | 249 | 0 |77503 | | KYE713L4 | LTO-4 |5 |360 | 151467 |169844 | 250 | 250 | 0 |77503 | | KYE713L4 | LTO-4 |6 |361 | 169844 |292514 | 251 | 251 | 0 |51961 | | KYE713L4 | LTO-4 |7 |913 | 1 |292514 | 0 | 251 | 0 |51961 | ++---+--+++---+---+-++--+ Well! Does the big gap in JobMediaId indicate that in fact this session was bscanned though it wasn't already purged? I tried this one: SELECT VolumeName, FirstIndex, JobId, count(*) FROM JobMedia, Media WHERE JobMedia.MediaId=Media.MediaId GROUP BY VolumeName, FirstIndex, JobId HAVING count(*) 1; which gave the following result: +++---+--+ | VolumeName | FirstIndex | JobId | count(*) | +++---+--+ | KYE713L4 | 1 |10 |2 | | KYE713L4 | 1 |13 |2 | | KYE713L4 | 1 |14 |2 | | KYE713L4 | 1 |15 |2 | [...] | KYE713L4 | 1 | 269 |2 | | KYE713L4 | 1 | 270 |2 | | KYE713L4 | 1 | 271 |2 | | KYE714L4 | 1 | 275 |2 | | KYE714L4 | 1 | 276 |2 | | KYE714L4 | 1 | 277 |2 | [...] | KYE714L4 | 1 | 361 |2 | | KYE714L4 | 1 | 362 |2 | | KYE714L4 | 1 | 363 |2 | | KYE714L4 | 1 | 364 |2 | | KYE714L4 | 1 | 368 |2 | | KYE715L4 | 1357 | 346 | 12 | | KYE715L4 | 1358 | 346 |5 | | KYE715L4 | 1767 | 353 | 17 | When I try your query with one of the JobId between 10 and 368 I always get something like this: ++---+--+++---+---+-++--+ | VolumeName | MediaType | VolIndex | JobMediaId | FirstIndex | LastIndex | StartFile | EndFile | StartBlock | EndBlock | ++---+--+++---+---+-++--+ | KYE714L4 | LTO-4 |1 |738 | 1 | 1 | 67 | 67 | 0 | 4982 | | KYE714L4 | LTO-4 |2 | 1007 | 1 | 1 | 0 | 67 | 0 | 4982 | ++---+--+++---+---+-++--+ Deleting the extra entries won't be a problem. Thanks a lot for putting me in the right direction! My problem is solved, but there seems to be an error somewhere in the bscan routines IIRC I used bscan -msS /dev/nst1 -V KYE713L4\|KYE714L4 Greetings, Michael -- The Palm PDK Hot Apps Program offers developers who use the Plug-In Development Kit to bring their C/C++ apps to Palm for a share of $1 Million in cash or HP Products. Visit us here for more details: http://p.sf.net/sfu/dev2dev-palm ___ Bacula-users mailing list
Re: [Bacula-users] Bacula restore runs forever - bootstrap creation [solved]
On Thu, 5 Aug 2010 11:46:27 +0200, news only said: Martin Simmons schrieb: What is the size of the job in bytes? exactly 28321715406 B (approx 27 GB) OK, that is very close to the total number of 64512 byte blocks recorded in VolIndex 1 to 6, so I think those rows represent the whole job (nothing missing or extra). ++---+--+++---+---+-++--+ | VolumeName | MediaType | VolIndex | JobMediaId | FirstIndex | LastIndex | StartFile | EndFile | StartBlock | EndBlock | ++---+--+++---+---+-++--+ | KYE713L4 | LTO-4 |1 |356 | 1 |101610 | 246 | 246 | 0 |77503 | | KYE713L4 | LTO-4 |2 |357 | 101610 |116651 | 247 | 247 | 0 |77503 | | KYE713L4 | LTO-4 |3 |358 | 116651 |126690 | 248 | 248 | 0 |77503 | | KYE713L4 | LTO-4 |4 |359 | 126690 |151467 | 249 | 249 | 0 |77503 | | KYE713L4 | LTO-4 |5 |360 | 151467 |169844 | 250 | 250 | 0 |77503 | | KYE713L4 | LTO-4 |6 |361 | 169844 |292514 | 251 | 251 | 0 |51961 | | KYE713L4 | LTO-4 |7 |913 | 1 |292514 | 0 | 251 | 0 |51961 | ++---+--+++---+---+-++--+ Well! Does the big gap in JobMediaId indicate that in fact this session was bscanned though it wasn't already purged? Yes, it is very likely. It looks like bscan creates only one JobMedia record per job and its StartFile is always 0. Deleting the extra entries won't be a problem. Thanks a lot for putting me in the right direction! My problem is solved, but there seems to be an error somewhere in the bscan routines Bscan has some limitations, but they are not all documented (or known!). __Martin -- The Palm PDK Hot Apps Program offers developers who use the Plug-In Development Kit to bring their C/C++ apps to Palm for a share of $1 Million in cash or HP Products. Visit us here for more details: http://p.sf.net/sfu/dev2dev-palm ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Bacula restore runs forever - bootstrap creation
On Sat, 31 Jul 2010 12:01:40 +0200, news only said: Martin Simmons schrieb: Michael schrieb: So, what might be the cause for the creation of the erroneous bootstrap entry (VolAddr=0-1078036843257)? It could be catalog corruption. What does the following SQL query return (where NNN is the jobid containing that file)? SELECT VolumeName, MediaType, FirstIndex, LastIndex, StartFile, JobMedia.EndFile, StartBlock, JobMedia.EndBlock, Slot, StorageId, InChanger FROMJobMedia, Media WHERE JobMedia.JobId=NNN AND JobMedia.MediaId=Media.MediaId ORDER BY VolIndex,JobMediaId; Thanks for the reply. This is the outout from the query: ++---++---+---+-++--+--+---+---+ | VolumeName | MediaType | FirstIndex | LastIndex | StartFile | EndFile | StartBlock | EndBlock | Slot | StorageId | InChanger | ++---++---+---+-++--+--+---+---+ | KYE713L4 | LTO-4 | 1 |101610 | 246 | 246 | 0 |77503 |1 | 3 | 1 | | KYE713L4 | LTO-4 | 101610 |116651 | 247 | 247 | 0 |77503 |1 | 3 | 1 | | KYE713L4 | LTO-4 | 116651 |126690 | 248 | 248 | 0 |77503 |1 | 3 | 1 | | KYE713L4 | LTO-4 | 126690 |151467 | 249 | 249 | 0 |77503 |1 | 3 | 1 | | KYE713L4 | LTO-4 | 151467 |169844 | 250 | 250 | 0 |77503 |1 | 3 | 1 | | KYE713L4 | LTO-4 | 169844 |292514 | 251 | 251 | 0 |51961 |1 | 3 | 1 | | KYE713L4 | LTO-4 | 1 |292514 | 0 | 251 | 0 |51961 |1 | 3 | 1 | ++---++---+---+-++--+--+---+---+ Ok, that explains the duplicate bootstrap rows. I do not understand the reason of the last line. Shouldn't it be StartFile 246? But I get similar results for at least another full backup (didn't query all of them). No, I don't think StartFile 246 would be any better. The last row looks completely bogus to me -- either it is not needed or it is badly corrupted. What is the size of the job in bytes? Did it complete without errors? Have you run any tools like bscan with this volume to update the catalog? What does the following slightly modified query print for that jobid? SELECT VolumeName, MediaType, VolIndex, JobMediaId, FirstIndex, LastIndex, StartFile, JobMedia.EndFile, StartBlock, JobMedia.EndBlock FROMJobMedia, Media WHERE JobMedia.JobId=NNN AND JobMedia.MediaId=Media.MediaId ORDER BY VolIndex,JobMediaId; __Martin -- The Palm PDK Hot Apps Program offers developers who use the Plug-In Development Kit to bring their C/C++ apps to Palm for a share of $1 Million in cash or HP Products. Visit us here for more details: http://p.sf.net/sfu/dev2dev-palm ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Bacula restore runs forever - bootstrap creation
Martin Simmons schrieb: Michael schrieb: So, what might be the cause for the creation of the erroneous bootstrap entry (VolAddr=0-1078036843257)? It could be catalog corruption. What does the following SQL query return (where NNN is the jobid containing that file)? SELECT VolumeName, MediaType, FirstIndex, LastIndex, StartFile, JobMedia.EndFile, StartBlock, JobMedia.EndBlock, Slot, StorageId, InChanger FROMJobMedia, Media WHERE JobMedia.JobId=NNN AND JobMedia.MediaId=Media.MediaId ORDER BY VolIndex,JobMediaId; Thanks for the reply. This is the outout from the query: ++---++---+---+-++--+--+---+---+ | VolumeName | MediaType | FirstIndex | LastIndex | StartFile | EndFile | StartBlock | EndBlock | Slot | StorageId | InChanger | ++---++---+---+-++--+--+---+---+ | KYE713L4 | LTO-4 | 1 |101610 | 246 | 246 | 0 |77503 |1 | 3 | 1 | | KYE713L4 | LTO-4 | 101610 |116651 | 247 | 247 | 0 |77503 |1 | 3 | 1 | | KYE713L4 | LTO-4 | 116651 |126690 | 248 | 248 | 0 |77503 |1 | 3 | 1 | | KYE713L4 | LTO-4 | 126690 |151467 | 249 | 249 | 0 |77503 |1 | 3 | 1 | | KYE713L4 | LTO-4 | 151467 |169844 | 250 | 250 | 0 |77503 |1 | 3 | 1 | | KYE713L4 | LTO-4 | 169844 |292514 | 251 | 251 | 0 |51961 |1 | 3 | 1 | | KYE713L4 | LTO-4 | 1 |292514 | 0 | 251 | 0 |51961 |1 | 3 | 1 | ++---++---+---+-++--+--+---+---+ I do not understand the reason of the last line. Shouldn't it be StartFile 246? But I get similar results for at least another full backup (didn't query all of them). Greetings Michael -- The Palm PDK Hot Apps Program offers developers who use the Plug-In Development Kit to bring their C/C++ apps to Palm for a share of $1 Million in cash or HP Products. Visit us here for more details: http://p.sf.net/sfu/dev2dev-palm ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Bacula restore runs forever - bootstrap creation
On Fri, 30 Jul 2010 11:06:30 +0200, news only said: I tried to restore one single file from a LTO-4 drive installed into a Quantum SuperLoader3 changer device. The file in question was just a small (15 kB) one. It was positioned somewhere in the middle of a tape. The tape contained about 200 backup sessions in about 280 tape-files. About 150 jobs where already purged, the file to be restored was in a session not purged. The tape contained approx. 800 GB and was in Error-Condition after a crash some time in the past. I use bacula version 5.0.1 with MySQL on SuSE Enterprise 9. When I did the restore I got the following bootstrap-file: -8-- Storage=SuperLoader3 Volume=KYE713L4 MediaType=LTO-4 Device=LTO-4 Slot=1 VolSessionId=129 VolSessionTime=1275651480 VolAddr=1060856922112-1060856999615 FileIndex=111445 Count=1 Storage=SuperLoader3 Volume=KYE713L4 MediaType=LTO-4 Device=LTO-4 Slot=1 VolSessionId=129 VolSessionTime=1275651480 VolAddr=0-1078036843257 FileIndex=111445 Count=1 -8-- The restore run forever and I cancelled it. ... So, what might be the cause for the creation of the erroneous bootstrap entry (VolAddr=0-1078036843257)? It could be catalog corruption. What does the following SQL query return (where NNN is the jobid containing that file)? SELECT VolumeName, MediaType, FirstIndex, LastIndex, StartFile, JobMedia.EndFile, StartBlock, JobMedia.EndBlock, Slot, StorageId, InChanger FROMJobMedia, Media WHERE JobMedia.JobId=NNN AND JobMedia.MediaId=Media.MediaId ORDER BY VolIndex,JobMediaId; __Martin -- The Palm PDK Hot Apps Program offers developers who use the Plug-In Development Kit to bring their C/C++ apps to Palm for a share of $1 Million in cash or HP Products. Visit us here for more details: http://p.sf.net/sfu/dev2dev-palm ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users