On Tue, 24 Dec 2024 10:32:13 -0500
"Brian J. Murrell via FreeIPA-users"
<[email protected]> wrote:

> On Fri, 2024-12-20 at 12:26 -0500, Brian J. Murrell via FreeIPA-users
> wrote:
> > In fact, checking on it again just now it seems to be even worse:
> > 
> > …
> > 
> > As you can see, it's a massive sustained write and tps rate and now
> > maxing out the disk mostly constantly.  This cannot be normal in any
> > circumstance, never mind a tiny little installation such as I have,
> > can
> > it?  
> 
> Since this issue seems to be eluding folks here, any suggestions on
> where else I should take this issue that people might have more insight
> or be able to suggest how to debug more deeply?  I really need this
> disk hammering to stop.  I cannot dedicate an entire disk to this
> process just because it wants to hammer away at a disk and since nobody
> else seems to be reporting the same I have to believe that my system is
> anomalous in this regard.
> 
> Cheers,
> b.
> 

I'm just a user so I might not be able to help you much, I left it to
the more experienced devs which know more about it, but what I would do:

From the 389-ds access log you sent, you can see the IP of the client
that's generating the requests (10.75.22.247)
What's weird to me is that I looked at my own server and I couldn't
find the same SRCH bases as you have in my access logs for the past 10
days, and I have about 20 hosts enrolled and alive. 
The particular SRCH strings I looked for are:
'replication agreement'
'replication changelog'
'pbac'
couldn't find them in my logs.
My hunch is it's as if something is continuously scanning your
directory. Maybe something malicious trying to download all accessible
data. I'd first confirm that the IP connecting is one of your own. Then
connect to that IP and tcpdump for connections to port 389 or 686 of
your IPA server (389-ds ports). I think only sssd should be initiating
connections to LDAP. Maybe also certmonger. Try shutting down sssd
(and/or certmonger) on that client and see if load goes down. If still
no, look at the IPs of clients connecting in 389-ds access log again,
and shut that client down, repeat until you bring the load down to
normal, which is almost 0 at least on my system. Then you know which
client(s) are causing the issue. Then increase debug_level in sssd.conf
(read sssd.conf man page) gradually and look at the client's sssd logs.
Maybe you can track down what is causing the excessive requests on the
client. What exactly to look for I don't know but maybe you will have
some clues there.

If the IP is of a IPA server or replica, that's a whole different can
of worms I don't know much about. There are lots of things on a IPA
server that connect to LDAP, you can try shutting down each service one
by one (httpd, pki-tomcat, named, ...)

Attachment: pgpfN4wfkiiQx.pgp
Description: OpenPGP digital signature

-- 
_______________________________________________
FreeIPA-users mailing list -- [email protected]
To unsubscribe send an email to [email protected]
Fedora Code of Conduct: 
https://docs.fedoraproject.org/en-US/project/code-of-conduct/
List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
List Archives: 
https://lists.fedorahosted.org/archives/list/[email protected]
Do not reply to spam, report it: 
https://pagure.io/fedora-infrastructure/new_issue

Reply via email to