Re: assertion failed: store_dir_diskd.c:839: "rb->flags.need_to_validate"

From: Andres Kroonmaa <andre@dont-contact.us>
Date: Fri, 6 Jul 2001 21:31:26 +0200

 More bits.
 I've noticed that sometimes swap.state is too large, say 3 times larger
 than expected. I suspect that Squid is in some conditions appending to
 swap.state instead of writing out clean and then rename.

 Can trigger this if during CLEAN startup send HUP signal to Squid. Then
 Squid stops writing swap.state.new, and instead continues to append stuff
 to swap.state. But this is not exact case, as after completing it simply
 renames unfinished swap.state.new to swap.state thus truncating swap.state
 badly. Thats another issue.

On 6 Jul 2001, at 19:07, Andres Kroonmaa <andre@online.ee> wrote:

> In short, 2.4 and latest HEAD fails to rebuild from CLEAN swaplogs when
> using diskd, and on Solaris. Seems that on Linux same compile works,
> although this seems very strange.
>
> Looking into code makes me think that this condition can only occur if
> during cleanlog write during earlier shutdown squid managed to write
> out bad cleanlog.
>
> Or, while reading cleanlog during startup, that log is appended with
> new objects.
>
> I'm not sure which one it is (I tend to think the latter), and how to
> fix it.

------------------------------------
 Andres Kroonmaa <andre@online.ee>
 CTO, Microlink Online
 Tel: 6501 731, Fax: 6501 725
 Pärnu mnt. 158, Tallinn,
 11317 Estonia
Received on Fri Jul 06 2001 - 13:37:03 MDT

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:14:06 MST