[OpenAFS-devel] Salvage extensive logging might be a problem later.

[Horst Birthelmer]

On Nov 15, 2004, at 12:44 PM, Harald Barth wrote:

>
> Yes, it is nice to know the progress of the salvager but I don't need
> to know it 100 times per second. Could this be limited to write
> something every 10 seconds or if that is not so good (because you
> don't want to make a syscall to get the time) every 1000 or 10000
> inodes?
>
> 11/15/2004 12:32:05 14617 inodes to process
> ....
> 11/15/2004 12:32:05 14524 inodes to process
>
> dweller:logs# grep 12:32:05 SalvageLog.0 | wc
>      94     564    4136
> dweller:logs# wc SalvageLog
>   52129  312833 2272238 SalvageLog
>
> This is a small test server, we have servers with many volumes
> containing > 100000 files each.

Yes, I made that change ...

I had some trouble with hanging salvaging.
If you used my AIX port you had that, too, without noticing :-)

Just comment that line out (It's only in the cvs version) never meant 
to enter the stable state.
And by the way, it's executed only when inodes are salvaged.

Horst