[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Problem with copy files >9 GB into OpenAFS



Am Freitag, 22. März 2013 schrieb Kevin Krammer:
> On Friday, 2013-03-22, Martin Steigerwald wrote:
> > Am Freitag, 22. März 2013 schrieb Lars Schimmer:
> > > Now if I try to copy a file >9GB (to be more exact: over the df
> > > reported by OpenAFS for /afs/) into a subdirectoy with enough quota,
> > > KDE copy breaks with a disk full error, although enough space is
> > > available.
> > > 
> > > KDE on squeeze did not have had this error/problem.
> > 
> > Does the error come up immediately, i.e. before it tries to copy
> > anything?
> > 
> > That would hint at the software checking reported free disk space. I
> > have never then this with KDE software so far and also for BTRFS the
> > reported value does not need to have much to do with reality.
> 
> How does on check for actually available space on those file systems?
> It is always problematic if etablished system interfaces start reporting
> bogus values and are replaced by newer interfaces that report the
> correct ones instead. But if the newer interfaces are pointed out then
> this could be taken care of.

I agree that this is problematic for applications.

But in BTRFS this has technical reasons:

https://btrfs.wiki.kernel.org/index.php/FAQ#Why_are_there_so_many_ways_to_check_the_amount_of_free_space.3F

https://btrfs.wiki.kernel.org/index.php/FAQ#Why_is_free_space_so_complicated.3F

Basically: Its too flexible for correct free space display.

And then this leads to something like:

merkaba:[…]/btrfs-progs-unstable> ./btrfs fi df /home
Disk size:               223.52GB
Disk allocated:          208.02GB
Disk unallocated:         15.50GB
Used:                    197.13GB
Free (Estimated):         26.38GB       (Max: 26.38GB, min: 18.63GB)
Data to disk ratio:         100 %

Yeah, estimated. While df -hT reports:

merkaba:~> LANG=C df -hT /home
Filesystem     Type   Size  Used Avail Use% Mounted on
/dev/dm-2      btrfs  224G  198G   25G  89% /home


(Note that is not yet into released BTRFS tools but in

git http://cassiopea.homelinux.net/git/btrfs-progs-unstable.git df-du-raid56
)

> > > Which package do we need to report a rc bug against?
> > > Thank you.
> > 
> > With what software did you copy the file? Dolphin? Then thats either
> > dolphin package or some core KDE dependency of it. I am not sure of
> > which.
> 
> Try with: kioclient copy source destination
> If that works it is something Dolphin does (or one of its libraries), if
> it doesn't then it is something KIO does (libkio, kdelibs).

It just fails silently with return code 1:

martin@merkaba:~/Zeit> truncate -s 100G 100g
martin@merkaba:~/Zeit> kioclient copy 100g 100g-2        
martin@merkaba:~/Zeit#1> ls -l 100g*
-rw-r--r-- 1 martin martin 107374182400 Mär 22 13:03 100g


Again with everything activiated in kdebugdialog (usually completely off
here):

martin@merkaba:~/Zeit> kioclient copy 100g 100g-2
unnamed app(17240) ClientApp::doIt: Creating ClientApp
kioclient(17240) KSharedUiServerProxy::KSharedUiServerProxy: kuiserver registered
kioclient(17240)/kio (Scheduler) KIO::SchedulerPrivate::doJob: KIO::SimpleJob(0x8d9070)
kioclient(17240)/kio (Scheduler) KIO::SchedulerPrivate::protoQ: creating ProtoQueue instance for "file"
kioclient(17240)/kdecore (KSycoca) KSycocaPrivate::openDatabase: Trying to open ksycoca from "/var/tmp/kdecache-
martin/ksycoca4"
kioclient(17240)/kio (Scheduler) KIO::ProtoQueue::ProtoQueue: m_maxConnectionsTotal: 5 m_maxConnectionsPerHost: 5
kioclient(17240)/kio (Slave) KIO::Slave::createSlave: createSlave "file" for KUrl("file:///home/martin/Zeit/100g-2")
kioclient(17240)/kio (KIOConnection) KIO::ConnectionServer::listenForRemote: Listening on  "local:/tmp/ksocket-
martin/kioclientw17240.slave-socket"
kioclient(17240)/kio (KIOJob) KIO::SlaveInterface::dispatch: error  111   "/home/martin/Zeit/100g-2"
kioclient(17240)/kio (Scheduler) KIO::SchedulerPrivate::jobFinished: KIO::StatJob(0x8d9070) KIO::Slave(0x8f66c0)
kioclient(17240)/kio (KDirWatch) KDirWatchPrivate::KDirWatchPrivate: Can't use FAM (fam daemon not running?)
kioclient(17240)/kio (KDirWatch) KDirWatchPrivate::KDirWatchPrivate: INotify available:  true
kioclient(17240)/kio (Scheduler) KIO::SchedulerPrivate::doJob: KIO::SimpleJob(0x8a63f0)
kioclient(17240)/kio (Scheduler) KIO::SchedulerPrivate::jobFinished: KIO::StatJob(0x8a63f0) KIO::Slave(0x8f66c0)
kioclient(17240)/kio (KIOJob) KIO::CopyJobPrivate::statCurrentSrc: Stating finished. To copy: 107374182400 , available: 
26655023104
kioclient(17240)/kio (Slave) KIO::Slave::kill: killing slave pid 17241 ( "file://" )
martin@merkaba:~/Zeit#1> ls -l 100g*               
-rw-r--r-- 1 martin martin 107374182400 Mär 22 13:03 100g
martin@merkaba:~/Zeit>


I don´t see why exactly it is failing here.

Thanks,
-- 
Martin 'Helios' Steigerwald - http://www.Lichtvoll.de
GPG: 03B0 0D6C 0040 0710 4AFA  B82F 991B EAAC A599 84C7


Reply to: