RAM-Cluster: Different filter sections on faceted search
Hi all,
After upgrade my RAM version, I experienced a strange behavior. A cluster instance show me all filters on the faceted search (Category, Type, State, Community) but the another not, especifically the State's filter does not appear on the mentioned search.
I've already rebuilt the index, reset the cache, remove the WAS cache, but without success, my next action will be delete the index folder content, forcing the phisical rebuilding.
What is the correct action in this case? Any other idea about what happens?
After upgrade my RAM version, I experienced a strange behavior. A cluster instance show me all filters on the faceted search (Category, Type, State, Community) but the another not, especifically the State's filter does not appear on the mentioned search.
I've already rebuilt the index, reset the cache, remove the WAS cache, but without success, my next action will be delete the index folder content, forcing the phisical rebuilding.
What is the correct action in this case? Any other idea about what happens?
7 answers
Hi
These are two different problems. For both issues I'm assuming that your servers are running on separate nodes.
1. Theme
In a cluster environment you must update the theme on each individual node. The theme does not automatically propagate to all nodes.
2. Filters
I'm not sure what is causing this problem. If you submit a new asset are you able to search for it on both servers? Could you perform a search for * on both servers and see if the number of assets match? Also, what version of RAM are you running?
These are two different problems. For both issues I'm assuming that your servers are running on separate nodes.
1. Theme
In a cluster environment you must update the theme on each individual node. The theme does not automatically propagate to all nodes.
2. Filters
I'm not sure what is causing this problem. If you submit a new asset are you able to search for it on both servers? Could you perform a search for * on both servers and see if the number of assets match? Also, what version of RAM are you running?
Hi Sheehan
My RAM is 7.5.0.2, running in a AIX 6, the cluster is design like this:
Node #1, running RAM and RTC (active)
Node #2, running RAM and RTC (stand-by)
I've only 2 servers running RAM, so it's not so difficult to meet to desired server and publish my own theme, but isn't a task to a non-technical administrator, and if were 4, 5 or more servers, it would be a true headache.
I submitted a new asset, as a result, on node that I sent my asset after 25 minutes it isn't on search results, but on the other node, it's available searching with * criteria after a 5 minutes.
On node "A" there are 313 assets and on the node "B" only 134.
I saw that some java core, heap and snap were created at the same time that I identified these problems, the OutOfMemory exception was logged on the java.core. Now the state filter on the faceted search go back to appears but with these labels: mylifecyle.state.s1, mylifecyle.state.s1, ..., mylifecyle.state.N.
My RAM is 7.5.0.2, running in a AIX 6, the cluster is design like this:
Node #1, running RAM and RTC (active)
Node #2, running RAM and RTC (stand-by)
1. Theme
In a cluster environment you must update the theme on each individual node. The theme does not automatically propagate to all nodes.
I've only 2 servers running RAM, so it's not so difficult to meet to desired server and publish my own theme, but isn't a task to a non-technical administrator, and if were 4, 5 or more servers, it would be a true headache.
2. Filters
I'm not sure what is causing this problem. If you submit a new asset are you able to search for it on both servers? Could you perform a search for * on both servers and see if the number of assets match? Also, what version of RAM are you running?
I submitted a new asset, as a result, on node that I sent my asset after 25 minutes it isn't on search results, but on the other node, it's available searching with * criteria after a 5 minutes.
On node "A" there are 313 assets and on the node "B" only 134.
I saw that some java core, heap and snap were created at the same time that I identified these problems, the OutOfMemory exception was logged on the java.core. Now the state filter on the faceted search go back to appears but with these labels: mylifecyle.state.s1, mylifecyle.state.s1, ..., mylifecyle.state.N.
On 3/28/2011 9:38 AM, rramoss wrote:
This is not a supported configuration. If both nodes are running at the
same time RTC will have problems. It does not support running on more
than one server at a time. Even if you say "on standby." It is still
considered active.
--
Rich Kulp
Rational Asset Manager developer
Hi Sheehan
My RAM is 7.5.0.2, running in a AIX 6, the cluster is design like
this:
Node #1, running RAM and RTC (active)
Node #2, running RAM and RTC (stand-by)
This is not a supported configuration. If both nodes are running at the
same time RTC will have problems. It does not support running on more
than one server at a time. Even if you say "on standby." It is still
considered active.
--
Rich Kulp
Rational Asset Manager developer
Hi Richkulp,
So, if two rtc are running at the same time, not as a cluster (obviously), and the RAM configuration setting to RTC "active" adress, I can experience this kind of trouble?
I've to keep the standby RTC down, and in case of failure to turn on it and reconfigure the RAM configuration page?
So, if two rtc are running at the same time, not as a cluster (obviously), and the RAM configuration setting to RTC "active" adress, I can experience this kind of trouble?
I've to keep the standby RTC down, and in case of failure to turn on it and reconfigure the RAM configuration page?
On 3/28/2011 11:38 AM, rramoss wrote:
Correct, they are considered to be two different RTC's.
However, RTC has a HA (High Availability) option. You should look into
that instead. HA is specifically tailored to what you are trying to do.
--
Rich Kulp
Rational Asset Manager developer
I've to keep the standby RTC down, and in case of failure to turn on
it and reconfigure the RAM configuration page?
Correct, they are considered to be two different RTC's.
However, RTC has a HA (High Availability) option. You should look into
that instead. HA is specifically tailored to what you are trying to do.
--
Rich Kulp
Rational Asset Manager developer
The HA option proposed was not accept by the client's support team, so the minor impact understood was to change the configuration in case of failure.
But it's not compreensive the "standby" RTC to be intrusive on the environment of execution. Changes on configuration doesn't happen every time, anyone access this server directly.
Anyway, we will stop the second RTC server and monitoring the behavior.
But it's not compreensive the "standby" RTC to be intrusive on the environment of execution. Changes on configuration doesn't happen every time, anyone access this server directly.
Anyway, we will stop the second RTC server and monitoring the behavior.