On Wednesday 19 July 2006 10:49, Sebastian Stark wrote:
> On 18.07.2006, at 18:47, Alan Brown wrote:
> > On Tue, 18 Jul 2006, Sebastian Stark wrote:
> >> I also noticed problems with this kind of setup but almost never
> >> got an
> >> answer when asking questions about multiple drive autochanger
> >> issues...
> >
> > For the most part, this is "esoteric high end use" :-)
>
> Probably. One of baculas strengths however is great scalability
> (we're at 50TB of backup data at the moment), that's why we use it in
> favour of other solutions :)

Yes, that has been one of the design goals.  For the next release, one of my 
focuses will be to document how to ensure that performance is satisfactory 
and discuss a bit about scaling for very large backups.

>
> >> ps I have three drives w/ thre concurrent jobs and spooling and
> >> very often
> >> find two of the drives idleing while there are hundreds of jobs
> >> waiting for a
> >> free drive.

Well, the above sounds more like a reservation system problem than a spooling 
problem.  The reservation system is *far* from optimal, and I wouldn't be 
surprised to find several more bugs that cause this kind of behavior, 
especially in this kind of "intensive" use situation, which is very hard to 
test.

> >
> > I am seeing this too.
> >
> > Additionally I'm getting multiple-hour hangs when trying "status
> > director"
> > in bconsole as it tries to check for jobs waiting to reserve a drive.



> >
> > The same hangs occur in the tray monitor.
> >
> >
> > Having observed it for a while, it looks like anything using drive 1
> > causes drive 0 to be locked out, but using drive 0 does not lockout
> > drive
> > 1.
>
> Sometimes it looks to me like bacula starts three jobs, then waits
> until *all* of them have finished and starts another round of three.
> I am not 100% sure yet, but I guess the other jobs wait "on max
> concurrent storage jobs", which I set to four. I have to find a way
> to monitor all this.
>
> I did not choose different spool directories for each drive, in my
> opinion one would only benefit from this by having a dedicated hard
> drive for each tape drive. Am I wrong?
>
> I have yet to find a way to reproduce this behaviour. It's a bit hard
> because you only notice this kind of stuff if you have relatively
> long running jobs. What I can do is send my configuration so far.
>
> bacula-sd.conf:
>
> Storage {
>    Name = yangtse-sd
>    SDPort = 9103
>    WorkingDirectory = "/usr/local/bacula/working"
>    Pid Directory = "/var/run"
>    Maximum Concurrent Jobs = 4
> }
>
> Autochanger {
>    Name = neo4000
>    Device = Drive3,Drive4,Drive5
>    Changer Command = "/usr/local/bacula/etc/mtx-changer %c %o %S %a %d"
>    Changer Device = "/dev/scsi/changer/c4t2100001086119230d2"
> }
>
> Device {
> # LTO-2 Laufwerk SCSI (NEO: Drive 1)
> # Used only for permanent pool (LTO-2 tapes only!)
> # thus autoselect=no
> Name = Drive1
> Media Type = LTO-2
> Archive Device = /dev/rmt/2cbn
> AutomaticMount = yes;
> AlwaysOpen = yes;
> RemovableMedia = yes;
> Drive Index = 0
> AutoChanger = yes
> Autoselect = no
> Spool Directory = /export/altai0/spool
> Maximum Spool Size     = 429496729600 # 400G
> Maximum Job Spool Size = 214748364800 # 200G
> Maximum Rewind Wait = 600
> Maximum Open Wait = 30
> Maximum Changer Wait = 120
> }
>
> Device {
> # LTO-3 Laufwerk FC (NEO: Drive 3) (dmesg: st18)
> Name = Drive3
> Media Type = LTO-2
> Archive Device = /dev/rmt/4cbn
> AutomaticMount = yes;
> AlwaysOpen = yes;
> RemovableMedia = yes;
> Drive Index = 1
> AutoChanger = yes
> Spool Directory = /export/altai0/spool
> Maximum Spool Size     = 429496729600  # 400G
> Maximum Job Spool Size = 214748364800  # 200G
> Maximum Rewind Wait = 600
> Maximum Open Wait = 30
> Maximum Changer Wait = 120
> }
>
> Device {
> # LTO-3 Laufwerk FC (NEO: Drive 4) (dmesg: st17)
> Name = Drive4
> Media Type = LTO-2
> Archive Device = /dev/rmt/3cbn
> AutomaticMount = yes;
> AlwaysOpen = yes;
> RemovableMedia = yes;
> #RandomAccess = no;
> Drive Index = 2
> AutoChanger = yes
> Spool Directory = /export/altai0/spool
> Maximum Spool Size     = 429496729600  # 400G
> Maximum Job Spool Size = 214748364800  # 200G
> Maximum Rewind Wait = 600
> Maximum Open Wait = 30
> Maximum Changer Wait = 120
> }
>
> Device {
> # LTO-3 Laufwerk SCSI (NEO: Drive 5) (dmesg: st14)
> Name = Drive5
> Media Type = LTO-2
> Archive Device = /dev/rmt/0cbn
> AutomaticMount = yes;
> AlwaysOpen = yes;
> RemovableMedia = yes;
> Drive Index = 3
> AutoChanger = yes
> Spool Directory = /export/altai0/spool
> Maximum Spool Size     = 429496729600  # 400G
> Maximum Job Spool Size = 214748364800  # 200G
> Maximum Rewind Wait = 600
> Maximum Open Wait = 30
> Maximum Changer Wait = 120
> }
>
>
>
>
> Sebastian
>
>
>
>
>
>
> -------------------------------------------------------------------------
> Take Surveys. Earn Cash. Influence the Future of IT
> Join SourceForge.net's Techsay panel and you'll get the chance to share
> your opinions on IT & business topics through brief surveys -- and earn
> cash
> http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
> _______________________________________________
> Bacula-users mailing list
> Bacula-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/bacula-users

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys -- and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_______________________________________________
Bacula-users mailing list
Bacula-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users

Reply via email to