On 01/12/2013 12:49 AM, Rob Crittenden wrote:
Rob Crittenden wrote:
Petr Viktorin wrote:
On 01/07/2013 05:42 PM, Rob Crittenden wrote:
Petr Viktorin wrote:
On 01/07/2013 03:09 PM, Rob Crittenden wrote:
Petr Viktorin wrote:
Works for me, but I have some questions (this is an area I know
Can we be 100% sure these certs are always renewed together? Is
certmonger the only possible mechanism to update them?
You raise a good point. If though some mechanism someone replaces
these certs it will cause the script to fail. Some notification of
failure will be logged though, and of course, the certs won't be
One could conceivably manually renew one of these certificates. It is
probably a very remote possibility but it is non-zero.
Can we be sure certmonger always does the updates in parallel? If it
managed to update the audit cert before starting on the others, we'd
no CA restart for the others.
These all get issued at the same time so should expire at the same
as well (see problem above). The script will hang around for 10
waiting for the renewal to complete, then give up.
The certs might take different amounts of time to update, right?
Eventually, the expirations could go out of sync enough for it to
AFAICS, without proper locking we still get a race condition when the
other certs start being renewed some time (much less than 10 min)
the audit one:
(time axis goes down)
audit cert other cert
certmonger does renew .
post-renew script starts .
check state of other certs: OK .
. certmonger starts renew
certutil modifies NSS DB + certmonger modifies NSS DB == boom!
This can't happen because we count the # of expected certs and wait
until all are in MONITORING before continuing.
The problem is that they're also in MONITORING before the whole renewal
starts. If the script happens to check just before the state changes
from MONITORING to GENERATING_CSR or whatever, we can get corruption.
The worse that would
happen is the trust wouldn't be set on the audit cert and dogtag
wouldn't be restarted.
The state the system would be in is this:
- audit cert trust not updated, so next restart of CA will fail
- CA is not restarted so will not use updated certificates
And anyway, why does certmonger do renewals in parallel? It seems
if it did one at a time, always waiting until the post-renew
done, this patch wouldn't be necessary.
From what Nalin told me certmonger has some coarse locking such that
renewals in a the same NSS database are serialized. As you point
would be nice to extend this locking to the post renewal scripts. We
ask Nalin about it. That would fix the potential corruption issue.
still much nicer to not have to restart dogtag 4 times.
Well, three extra restarts every few years seems like a small price to
pay for robustness.
It is a bit of a problem though because the certs all renew within
seconds so end up fighting over who is restarting dogtag. This can
some renewals go into a failure state to be retried later. This is fine
functionally but makes QE a bit of a pain. You then have to make sure
that renewal is basically done, then restart certmonger and check
everything again, over and over until all the certs are renewed.
difficult to automate.
So we need to extend the certmonger lock, and wait until Dogtag is back
up before exiting the script. That way it'd still take longer than 1
restart, but all the renews should succeed.
Right, but older dogtag versions don't have the handy servlet to tell
that the service is actually up and responding. So it is difficult to
tell from tomcat alone whether the CA is actually up and handling
Revised patch that takes advantage of new version of certmonger.
certmonger-0.65 adds locking from the time renewal begins to the end of
the post_save_command. This lets us be sure that no other certmonger
renewals will have the NSS database open in read-write mode.
We need to be sure that tomcat is shut down before we let certmonger
save the certificate to the NSS database because dogtag opens its
database read/write and two writers can cause corruption.
stop_pkicad and start_pkicad need the Dogtag version check to select
A more serious issue is that stop_pkicad needs to be installed on
upgrades. Currently the whole enable_certificate_renewal step in
ipa-upgradeconfig is skipped if it was done before.
In stop_pkicad can you change the first log message to "certmonger
stopping %sd"? It's before the action so we don't want past tense.
Freeipa-devel mailing list