Ive got a problem with my QoS, since we moved our Oracle 9i server to the data centre where our internet connection is presented. A large download from the internet stops traffic getting to the Oracle server.
We have two 2621XM routers that link between the sites over a LES 10 circuit, and have applied QoS configuration to support LLQ of 55% on them. I have tried using different methods of matching, i.e. via access-lists to identify source/destination address, source/destination port, NBAR, etc.
As we use the standard Oracle listening port (1521), Ive just kept it simple and using the match protocol sqlnet to identify traffic.
I have noticed that the traffic is fairly minimal to the server, the login and selecting the databases to open is only a few packets. The actual opening of the database is a couple thousand packets.
The problem seems to happen when opening the database if a large download from the internet is happening at the same time. Rather than the database being displayed after round 20 seconds it just sits there for over 2 minutes, at this time we terminate the internet download to avoid causing users problems.
The policy map seems to indicate that QoS is operating as expected, i.e. there are packets in the default class being tail dropped and it is matching the Oracle traffic.
Can someone help me out as I have other QoS configurations on different sites that work perfectly.
In summary the config is;
class-map match-any MISSION-CRITICAL-DATA
match protocol sqlnet
priority percent 55
set dscp af31
shape average 10000000
service-policy output Shaper
LES 10= 172.16.10.0/30
Please find attached router config extracts and copy of the debug (sorry time between routers wasnt correct, when test was done)
Out of curiosity, on the MISSION-CRITICAL-SERVERS AClL, why have you specified the port (1521) as the destination port. If the servers are located at the GQ site, the port should be the source port. In that case, you can use the ACL as your class-map match criteria.
The ACL matching with class-maps is pretty reliable. I suggest that you use a sniffer like Ethereal on the LAN segment where your server is and try and see the actual port numbers used by it. Then, you can have an ACL that more accurately matches the server traffic.
Many thanks for you assistance, I have now modified the policy map to match via an ACL using either source/destination address. (after examining the listener.ora file it was determined the box could listen on both the default 1521 & 1526).Also I changed the QoS from LLQ (priority percent 55) to 'Bandwidth 80 percent' this seems to bring up the Oracle database, but can't check thoroughly until later today.
I will let you all know if this resolves my issue.
Question We run asr9001 with XR 6.1.3, and we have a very long delay to
login w/ SSH 1 or 2 to the device compare to IOS device. After
investigation, the there is 1s delay between the client KEXDH_INIT and
the server (XR) KEXDH_REPLY. After debug ssh serv...
Introduction The purpose of this document is to demonstrate the Open
Shortest Path First (OSPF) behavior when the V-bit (Virtual-link bit) is
present in a non-backbone area. The V-bit is signaled in Type-1 LSA only
if the router is the endpoint of one or ...
Hi, I am seeing quite a few issues with patch install and wanted to
share my experience and workaround to this. Login to admin via CLI, then
access root with the “shell” command Issue “df –h” and you’ll probably
see the following directory full or nearly ...