- CheckMates
- :
- Products
- :
- Quantum
- :
- Maestro Masters
- :
- Unreached OCSP - causes serious lag
- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Mute
- Printer Friendly Page
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Unreached OCSP - causes serious lag
Hi,
Setup
Current setup is a Maestro R81 Take12. (already the case on Take10)
MHO140 /9300&9400's.
These is a serious delay in browsing websites thru the SG's.
Nearly all Security features are on - Sandblast suite.
Symthom
When browsing to a domain, the initial delay is 12-15sec, before it starts loading.
Seen these kinds of messages before in R81.20, but to compare :
- R81.20 - 25msg in 1,5h, 500+ endpoints active 24/7
- R82 - 25msg in 1,5min 5active users.
So far
TAC is already involved, but wondering if somebody else is experiencing this issue?
Currently no big clue where the issue may reside.
The OCSP error very present, and near the timeline of the actions taken.
Certificate used in HTTPS Inspect is exactly the same as the existing setup.
It's, apart of a RemoteAccess configuration move, the last hurdle I need to fix before being able to go live.
Kind regards
Tim
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
OSCP is occurring because we are validating the certificate for the remote site is valid as part of SNI verification we do.
This occurs on anything where SNI is processed (App Control, URL Filtering, and HTTPS Inspection among others).
If the OSCP connection is failing somehow (you didn't include the message, but I assume that's what they are), that would account for the slowness of establishing the initial connection.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
OSCP would definitely have to do with certificate. Where is the error? Does it happen on client side or somewhere else?
Andy
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hi @TimV,
Is there a way you can switch to CRL for certificate validation and see if it improves your situation?
We had a scenario some time ago with validating VPN authentication certificates. From memory, there was a change in behaviour that defaulted to using OCSP, but when we switched back to CRL validation, it resolved our issue.
Thanks,
Aaron.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hello,
How you would you do this?
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
To resolve our issue, we used Scenario 2 in SK179434.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Never seen that sk before, GREAT reference @AaronCP
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Both issues have the same root cause, as explained earlier.
The Certificate used for HTTPS Inspection must be a Certificate Authority (i.e. can't be a wildcard cert).
The CRL/OSCP server used by the CA needs to be reachable by the gateway (also client in case of Remote Access).
Beyond that, no specific guidance.
