Re: [Bacula-users] Support for HAVE_POSIX_FADVISE on Bacula Client

2015-10-12 Thread Josh Fisher
On 10/10/2015 10:15 PM, Kern Sibbald wrote: > Hello, > > >From what I read on the web, this whole thing is a bit of a mess, and > apparently POSIX requires that O_ACCMODE not be 3 because there are at > least two bits besides O_RWRD and O_RDONLY. However, on most systems > O_ACCMODE is 3. It wo

Re: [Bacula-users] Support for HAVE_POSIX_FADVISE on Bacula Client

2015-10-10 Thread Kern Sibbald
Hello, >From what I read on the web, this whole thing is a bit of a mess, and apparently POSIX requires that O_ACCMODE not be 3 because there are at least two bits besides O_RWRD and O_RDONLY. However, on most systems O_ACCMODE is 3. It would be hard to consider any part of flags as an integer b

Re: [Bacula-users] Support for HAVE_POSIX_FADVISE on Bacula Client

2015-10-10 Thread Richard Hall
On 10/10/2015 19:30, Kern Sibbald wrote: > On 10/10/2015 10:16 AM, Phil Stracchino wrote: >> On 10/09/15 22:05, Kern Sibbald wrote: >>> Hello, >>> >>> Thanks for pointing this out. I still have problems imagining that they >>> would define O_RDONLY as 0 in a bitmapped variable!!! >> It makes more s

Re: [Bacula-users] Support for HAVE_POSIX_FADVISE on Bacula Client

2015-10-10 Thread Kern Sibbald
On 10/10/2015 10:16 AM, Phil Stracchino wrote: > On 10/09/15 22:05, Kern Sibbald wrote: >> Hello, >> >> Thanks for pointing this out. I still have problems imagining that they >> would define O_RDONLY as 0 in a bitmapped variable!!! > It makes more sense if you think of it as the absence of O_RDWR.

Re: [Bacula-users] Support for HAVE_POSIX_FADVISE on Bacula Client

2015-10-10 Thread Phil Stracchino
On 10/09/15 22:05, Kern Sibbald wrote: > Hello, > > Thanks for pointing this out. I still have problems imagining that they > would define O_RDONLY as 0 in a bitmapped variable!!! It makes more sense if you think of it as the absence of O_RDWR. -- Phil Stracchino Babylon Communications p

Re: [Bacula-users] Support for HAVE_POSIX_FADVISE on Bacula Client

2015-10-09 Thread Kern Sibbald
decide. The default should be the > current behavior to not call posix_fadvise(). Also the code path for > SD/Dir und FD should be separate so enable this for the FD bfile.c > code only (it seems currently it is the same code). > > > > Regards, > > Robert > >

Re: [Bacula-users] Support for HAVE_POSIX_FADVISE on Bacula Client

2015-10-08 Thread Eric Bollengier
Hello Robert, Le 08. 10. 15 13:54, Robert Heinzmann a écrit : > Hi, > >> I think you did a really nice investigation work :-) > Thanks ... was fun :) > >> I think that your patch is almost right, and we need to figure if we want to >> add specific directives for that. I tend to believe right now

Re: [Bacula-users] Support for HAVE_POSIX_FADVISE on Bacula Client

2015-10-08 Thread Robert Heinzmann
-Ursprüngliche Nachricht- Von: Eric Bollengier [mailto:eric.bolleng...@baculasystems.com] Gesendet: Mittwoch, 7. Oktober 2015 16:50 An: Robert Heinzmann Cc: bacula-users@lists.sourceforge.net Betreff: Re: [Bacula-users] Support for HAVE_POSIX_FADVISE on Bacula Client Hello Robert, I thi

Re: [Bacula-users] Support for HAVE_POSIX_FADVISE on Bacula Client

2015-10-08 Thread Josh Fisher
free-electrons.com/source/mm/fadvise.c#L115), so there is “right” >> way to fix this issue >> >> - Best approach: there should be a Bacula Option to enable the >> POSIX_FADV_NOREUSE explicitly for selective bacula Jobs to allow usage of >> posix_fadvise

Re: [Bacula-users] Support for HAVE_POSIX_FADVISE on Bacula Client

2015-10-07 Thread Eric Bollengier
be the current behavior to not call > posix_fadvise(). Also the code path for SD/Dir und FD should be separate so > enable this for the FD bfile.c code only (it seems currently it is the same > code). > >   > > Regards, > > Robert > >   > > Robert Heinzma

Re: [Bacula-users] Support for HAVE_POSIX_FADVISE on Bacula Client

2015-10-07 Thread Gary R. Schmidt
On 7/10/2015 10:57 PM, Robert Heinzmann wrote: > Hello, > > just a short followup. It seems I found the issue causing our file > system cache to fill up during backup jobs and our virtual platform > memory usage to increase. > > The O_RDONLY mask is “00” on Linux, casing the if flags & O_RDONLY to

Re: [Bacula-users] Support for HAVE_POSIX_FADVISE on Bacula Client

2015-10-07 Thread Robert Heinzmann
me code). Regards, Robert Robert Heinzmann r.heinzm...@freelancer.traviangames.com IT-Operations Travian Games GmbH Von: Robert Heinzmann [mailto:r.heinzm...@freelancer.traviangames.com] Gesendet: Mittwoch, 7. Oktober 2015 11:16 An: bacula-users@lists.sourceforge.net Betreff: [Bacula-users] Support fo

[Bacula-users] Support for HAVE_POSIX_FADVISE on Bacula Client

2015-10-07 Thread Robert Heinzmann
Hello, we are using bacula on a virtual plattform (> 600 clients) and investiate moderate swap usage issues / cache usage issues. It seems that during backup of mysql datafiles (LVM Snapshot), the "cached" memory usage of the server increases and thus bacula backup reads (only done one) are cac