RE: --enable-async-io code status

On Wed, 6 Jan 1999, Jeff Madison wrote:

> I noticed this message this morning and thought a problem that I am having
> may be of interest to you. Below is a message that I posted to the list
> last night and describes a problem that many who use -enable-async-io are
> encountering. I run Solaris 2.6_x86 with all patches. See the message
> below:

I'm running Squid 2.1PATCH1 on SPARC-based Solaris machines and they seem
fine with AIO.

> with -enable-async-io. With this option enable Squid appears to "forget" or
> loose track of what and how many items it has in it's cache directory. I

How did you track/notice this? (Maybe I have a problem and don't even realise
it.. ;-)

  Storage Swap size: 17176756 KB
  Mean Object Size: 11.89 KB

So I'd estimate around 1.4 million objects should be present.

Internal Data Structures:
        1447064 StoreEntries
         18988 StoreEntries with MemObjects
         17412 Hot Object Cache Items
        1445128 Filemap bits set
        1445064 on-disk objects

Seems okay..

> start Squid and it would say it only had approx 60000 objects in the cache
> directory, hard to believe when it's using 40 GB of drive space. After

From last night's log rotation:

1999/01/07 00:00:09| storeDirWriteCleanLogs: Starting...
[...]
1999/01/07 00:00:46| 1441792 entries written so far.
1999/01/07 00:00:46| Finished. Wrote 1443609 entries.
1999/01/07 00:00:46| Took 37 seconds (39016.5 entries/sec).

Still around 1.4 million.

> ever finishes rebuilding forcing a restart and restarting the rebuild. We
> are running Squid 2.0.Patch2. My question is has this bug been fixed and in
> what version or does anyone have any suggestions.

Don't know for certain, but I'd suggest giving 2.1PATCH2(++) a try.

Cheers..

dave
Received on Wed Jan 06 1999 - 23:17:30 MST