Hmm, that's too bad about your FC5 experience. I've been running it here for a couple of months without any of these issues on an AMD 64 machine. I also upgraded from FC4. Granted I don't have any printers attached to it all my printers are network attached.
As far as the compiler is concerned, I don't think that is really the fault of Fedora. The compiler folks have just improved? the optimization. One of the drawbacks of increased optimization is you need to make some assumptions about the code. While those assumptions may be valid with some code it isn't valid with others and the compiler tells you not to use those optimizations. Adding -fno-strict-aliasing should eliminate most of those warnings and make the code more stable. What Window manager are you using with X? I haven't had any problems with gnome but I don't think KDE is as well tested. The only issue I've ever been able to attribute directly to redhat was a change they made to mtx that is definitely busted and doesn't work with some tape drives but I think it was also in FC4. I'll take a look at the st driver changes between FC4 and FC5 and see what they've done. -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Kern Sibbald Sent: Tuesday, June 13, 2006 7:09 AM To: bacula-announce Cc: bacula-devel; bacula-users Subject: [Bacula-users] Bacula version 1.38.10 released to Source Forge Hello, [...Release Information Removed....] Fedora FC5 warning !!!!!!!!!!!!!!!!!! I have recently upgraded a number of my machines (fortunately not my web server) to FC5, and I can say without reservation that it is the most unstable RedHat/Fedora system that I have ever seen. Among my gripes: - The new GNU C++ compiler spews out tons of warning messages for Bacula mostly concerning "type punned pointers and strict anti-aliasing problems". It is now necessary to be a rocket scientist before even considering using the GNU C++ compiler :-( - If I attempt to configure my printer, cups consumes every byte of available space on my / partition continuously logging the same authentication error message. - All "lpr filename" output gets printed in Landscape rather than Portrait (programs such as OO and browsers that explicitly set the orientation do print correctly). - At random all my Window borders disappear requiring me to abort X otherwise there is NO window management. - The Bacula storage daemon gets a lot of errors of the following kind: Matou.2006-06-13_03.05.00 Error: block.c:538 Write error at 71:0 on device "HP DLT 80" (/dev/nst0). ERR=Device or resource busy. which is absurd; there is only one file descriptor open, how can it be busy on a write request? - While running the SD under gdb debugging this problem (which IMO is a kernel driver bug), the OS hangs. An OS hang while running gdb, you got to be kidding, I haven't seen that for 5 years -- cool guys! - The FC4 to FC5 upgrade fails in numerous ways, often leaving you with a system with half FC4 rpms and half FC5 rpms (this happened on several systems). Trying to upgrade a half FC4/FC5 system is not fun, because nearly every way you turn, you have unresolved dependencies ... In short FC5 is a *BAD* release, and if you are on FC4 and considering moving up to FC5, I would recommend reading over my list of woes before making any decision. Concerning the FC5 rpms. Scott is not yet running an FC5 system (good choice Scott), so I am going to make the FC5 rpms, but given the general lack of stability, I am not sure when or how I am going to do it. I'd sure like to get back to programming, and starting to wonder if it is possible to upgrade from FC5 to FC4 ... -- Best regards, Kern ("> /\ V_V [ReleaseNotes] I Release Notes for Bacula 1.38.10 Bacula code: Total files = 423 Total lines = 139,479 (*.h *.c *.in) Warning for version 1.38.8 and greater: - The Windows FD no longer automatically folds the case in wild-card comparions. To get the same behavior as before, you must explicitly use "Ignore Case = yes" in your FileSet. - Backslashes are no longer permitted in File directives (typical error for Windows users), unless the string is enclosed in double quotes, in which case, the backslashes must be doubled. Major Changes for 1.38.10: - This is a primarily a bug fix release, with the following bugs being resolved #570, 609, 611, 613, 614, 617 as well as other bugs not reported in the bugs database. See below for details. - The bconsole "reload" command should now work properly in virtually all situations thanks to the efforts of a number of people (Eric, Christopher, and myself). - There is an untested fix to eliminate the scheduling problems (either no job run or the same job run twice) during sudden clock shifts (daylight savings time change). - The manual has been released under the GNU Free Documentation License (FDL). The major difference is that this removes the prior restriction on commercial reproduction of the manual. - Fixed a race condition (thanks to Christopher) on dual processor machines that caused jobs to block when starting with the SD. New features for 1.38.10: - None Fixes for 1.38.10: 08Jun06 - Modify setting ACLs to suppress error message if the file is a symbolic link since the file linked to may not yet be created, and symbolic links really should not have attributes ... 06Jun06 - Complete implementation of acl.c error message during restore. During backups, errors getting acls are silently ignored, as was previously the case. 04Jun06 - Modify configure.in to check for typeof(). - Better detection of postgresql installation if pg_conf works. - Fix RescheduleTimes = 0 bug (should not reschedule). - Down port from 1.39: fixes for reload, autochanger error messages, scheduler fixes for reload and day light savings change, 01Jun06 - Implement partial error messages for src/filed/acl.c - Fix return code to fail when doing update barcodes and Pool resource is not found. This previously caused Volumes to be labeled as cleaning tapes. - Small updates to projects file. 26May06 - Prevent DVD code from rewriting label if the device cannot be opened read/write. - Fix Dir/SD run race in migrate.c, verify.c, and restore.c - Update projects file. 24May06 - Applied patch posted by Jaime Ventura to bug #570 to correct failure of restarted jobs to complete (due to a destroyed pthreads conditional variable used by the message thread). - Fixed bug #619 where Bacula would not restart jobs with MaxRestartTimes set to zero. - Apply patch from John Goerzen bug #611 to fix bad manual links in the code. - Apply patch from Rudolf Cejka bug #614 to removed trailing colons in query.sql file. - Apply patch from Rudolf Cejka bug #613 to correct prunning of files not to create orphans. - Apply patch from Rudolf Cejka bug #617 to use mtime instead of ctime in restore long listing. - Apply patch from Rudolf Cejka bug #609 to use MarkId in verify.c instead of incorrect MarkedId. - Apply patch from adioso bug #616 to correct text in manual (replace Director with File daemon). 21May06 - Move DIR "run" command to SD before starting the message thread as suggested by Cristopher Hull to avoid a race deadlock from two threads using the bsock structure. - Modify LICENSE to correct some problems pointed out by Debian. 05May06 - Correct EOF with no \n lex.c problem. - Add additional INFO messages if autochanger cannot be used. - Add Eric Bollengier's lex error handler fix, which corrects most of the reload problems. _______________________________________________ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users _______________________________________________ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
