On 12/31/18 12:43 PM, Phil Stracchino wrote: > On 12/31/18 8:56 AM, Martin Simmons wrote: >>>>>>> On Sun, 30 Dec 2018 00:08:25 -0500, Phil Stracchino said: >>> >>> This failure mode (which looks like a block number integer overflow in >>> the storage daemon on Solaris 11.3) is solidly repeatable in 9.4.1. I >>> have had to roll back to 9.2.2 for now pending a fix. >> >> It looks like EndBlock=%d should be EndBlock=%u in >> BDB::bdb_update_media_record (bacula/src/cats/sql_update.c line 378). >> >> This will be broken for both 64-bit and 32-bit. > > That would do it. Just patched that, recompiling now and will try it again.
Nope, that is not sufficient. I even tried making it %lu and I am still getting negative EndBlocks. The integer overflow appears to be happening earlier in the process. I see from cats.h that both JOBMEDIA_DBR and MEDIA_DBR have uint32_t EndBlock; Should that be a uint64_t ? -- Phil Stracchino Babylon Communications ph...@caerllewys.net p...@co.ordinate.org Landline: +1.603.293.8485 Mobile: +1.603.998.6958 _______________________________________________ Bacula-devel mailing list Bacula-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-devel
