cluster size incorrect - VxWorks
This is a discussion on cluster size incorrect - VxWorks ; Hi
I recently shifted my code from Vx5.5 to Vx6.6.
i am seeing a peculiar problem after making the switch. I have a FAT16
dos filesystem on a RamDrive. After creating some files on this file
system, if I run ...
-
cluster size incorrect
Hi
I recently shifted my code from Vx5.5 to Vx6.6.
i am seeing a peculiar problem after making the switch. I have a FAT16
dos filesystem on a RamDrive. After creating some files on this file
system, if I run chkdsk command, i get the following error for
multiple files,
"filename.xxx" file larger than its chain of clusters Entry will be
deleted.
I have moved the filsystem to different areas of the RAM just to make
sure that no other process is messing with this RAM area. The error
scenario is also very consistent.
Any clue on what could cause this behavior will be appreciated.
Thanks.
-
Re: cluster size incorrect
nsc wrote:
> Hi
>
> I recently shifted my code from Vx5.5 to Vx6.6.
>
> i am seeing a peculiar problem after making the switch. I have a FAT16
> dos filesystem on a RamDrive. After creating some files on this file
> system, if I run chkdsk command, i get the following error for
> multiple files,
>
>
> "filename.xxx" file larger than its chain of clusters Entry will be
> deleted.
>
> I have moved the filsystem to different areas of the RAM just to make
> sure that no other process is messing with this RAM area. The error
> scenario is also very consistent.
>
> Any clue on what could cause this behavior will be appreciated.
>
> Thanks.
>
You mean other than the fact that you appear to be using a bleeding edge
not yet released version of the OS?
(I thought 6.4 just came out)
In any case a few random guesses.
1) Files not closed yet and newer vxWorks using more aggressive
buffering than 5.5 (potentially different default values for
disk/cache/buffers and the DoSFS not really realizing that underlying
device is a ramdisk and thus does not really need the buffers)
2) Even if the files are closed, it could be possible that buffering is
still playing a role.
I'd open a TSR.