Glad to hear it. Ralf, maybe we can get a FAQ for this?
---
"Why do I get lots of random SSL errors under heavy load?"
There can be a number of reasons for this, but the main one is problems
with the SSL Session Cache specified by the SSLSessionCache. The dbm
session cache is most likely the source of the problem, so trying the shm
session cache or no cache at all may help.
---
Thanks,
Dave
On Wed, Jan 17, 2001 at 07:39:11PM -0700, Albert Wold wrote:
> Thanks for this suggestion! We finally got our production servers switched
> over to shm, and it seems to have completely solved the problem. There have
> been no errors since we made the change.
>
> -----Original Message-----
> From: David Rees [mailto:[EMAIL PROTECTED]]
> Sent: Thursday, January 11, 2001 11:12 AM
>
> > The lines from httpd.conf are:
> >
> > SSLSessionCache dbm:/usr/local/apache/logs/ssl_scache
> > SSLSessionCacheTimeout 300
> >
> > Should we maybe be using shm instead? I can imagine that dbm would have
> > trouble if there was a lot of concurrent access attempts.
>
> I would definately try the shm session cache, I found it to improve
> performance quite a bit over the dbm session cache.
>
> I'm currently using the shmcb session cache (vs the shmht or shm cache)
> which is supposed to be better under heavy load, but I haven't seen your
> problem when using either. To enable this you need to compile with
> the --enable-rule=SSL_EXPERIMENTAL when configuring mod_ssl and then specify
> SSLSessionCache shmcb:/path/to/ssl_scache(size of cache).
>
> > The server gets about 50,000 to 100,000+ page views per day, and there are
> > around 25-50 active sessions/logins in our web application during the day.
>
> Doesn't seem like too much load on Apache itself...
>
> > Also, the errors you said were normal...is that because the user
> > pressed the
> > Stop button? I don't think that is the case in this situation, because
> > there are a ton of them all in a row, which normally wouldn't
> > happen if the
> > users were pressing stop occasionally.
>
> Usually they'll push stop, (or back, or whatever) and there the page would
> have already started loading all the graphics/etc for the page, which is why
> you see a bunch of them at once. Either way, we see this on our servers
> with similar load and I don't know of any problems. But I am not seeing
> those other errors you showed us. We have had a few busier days with
> hundreds of user sessions and about a million hits (not page views) in a 24
> hour period with no problems, but this is on a SGI machine. We do have some
> machines using Redhat 6.2 and mod_ssl, but they only see very light load.
______________________________________________________________________
Apache Interface to OpenSSL (mod_ssl) www.modssl.org
User Support Mailing List [EMAIL PROTECTED]
Automated List Manager [EMAIL PROTECTED]
