mount: Not a directory

[Petr Balas]

Kaper wrote:

> 
> 
>>Proc to CDcko pred stahovanim mountujete? IMHO je to zbytecne/nesmysl.
>>Ale protoze to je stejne readonly, tak by to nemelo nicemu vadit.
> 
> Ja vim - ja to tam psal jen proto, aby bylo videt, ze samo o sobe je v
> poradku.
> pri tomto mountu je v logu toto:
> 
> Apr 24 01:02:46 alnagon kernel: hdc: ATAPI 32X DVD-ROM CD-R/RW drive,
> 2048kB Cache, DMA
> Apr 24 01:02:46 alnagon kernel: Uniform CD-ROM driver Revision: 3.12
> Apr 24 01:02:46 alnagon kernel: hdd: ATAPI 50X CD-ROM drive, 128kB
> Cache, (U)DMA
> Apr 24 01:02:46 alnagon kernel: VFS: Disk change detected on device
> ide1(22,64)
> Apr 24 01:02:46 alnagon kernel: ISO 9660 Extensions: Microsoft Joliet
> Level 3
> Apr 24 01:02:46 alnagon kernel: ISOFS: changing to secondary root
> 
> 
>>Zkontrolujte, zda ta velikost CDcka odpovida realite (pokud je nekde
>>skrabnute, tak se nemuselo stahnout cele)
> 
> To bude pro me asi trochu problem.. mam v tom trochu chaos, jestli
> jsou stejne to nepoznam, ale zjistil jsem toto:
> 
> alnagon:~# df /dev/hdd
> Filesystem           1k-blocks      Used Available Use% Mounted on
> /dev/hdd                545398    545398         0 100% /mnt/cdrom
> alnagon:~# du -s /mnt/cdrom/
> 542216  /mnt/cdrom

To vypada rozumne.


>   >Popr. zkuste na ten mount pustit strace a uvidite, co se mu nelibi.
>   Takovej expert zase nejsem, ale co by mohlo byt podezrele, tady
> predkladam:
> 
> open("/etc/ld.so.preload", O_RDONLY)    = -1 ENOENT (No such file or
> directory)
> ioctl(4, 0x4c00, 0x3)                   = -1 ENOSYS (Function not
> implemented)
> ioctl(4, 0x4c02, 0xbffff7b4)            = -1 EINVAL (Invalid argument)
> open("/proc/crypto/cipher/(null)-cbc", O_RDONLY|O_LARGEFILE) = -1
> ENOENT (No such file or directory)
> ioctl(4, 0x4c02, 0xbffff7b4)            = -1 EINVAL (Invalid argument)
> stat64("/sbin/mount.iso9660", 0xbffff7b8) = -1 ENOENT (No such file or
> directory)
> mount("/dev/loop0", "/mnt/image", "iso9660", MS_RDONLY|0xc0ed0000,
> 0x805c120) = -1 ENOTDIR (Not a directory)
> semget(32, 1074975744, 0)               = -1 ENOSYS (Function not
> implemented)

Takze to hlasi primo mount sluzba a tak se z toho nic neda domyslet.


>>Podla mna to je znamy readahead bug v kerneli. Aku mas verziu? Skus
> novsiu.
>  mam Debian Woody na 2.4.18 .. "readahead bug" ?.. co to je? to je
> mimo moje schopnosti. mohlo by to byt ale ucit se patchovat a muset
> kompilovat kernel za tohle nestoji, neda se to nejak slusne obejit?
> 
>>Popr. by pak mohlo stacit pridat na konec toho souboru par prazdnych
>>sektoru.
> 
>  Ano, to bych zkusil. Ale jak tedy maji ty sektory byt velke? neco jako
>  dd if=/dev/zero of=/usr/image.iso bs=1024 skip=550544 count=4
>  ???

Padding je tusim ze na 16 sectoru, takze bych jich pridal aspon 16 a
to po 2048 bajtech. Resp. kdyz jich pridate vic, tak se nic nestane.
Ja bych to delal prez:
dd if=/dev/zero of=/usr/image.tmp bs=2048 count=64
cat /usr/image.tmp >> /usr/image.iso
rm /usr/image.tmp
Za pokus to IMHO stoji.

-- 
Petr Balas (petr at balas dot cz)