cluster cache stops after two hours or less

Quote from cesarin04

hi there,

i have this problem after 2 or 3 hours my cache stops working and i will only recive F and C but i wont recive no more R so i have to reboot the csp to the r to work again and after teo hours the ame thing happen
any suggestions ?

<cache-handler class="com.bowman.cardserv.ClusteredCache">
<cache-config>
<cw-max-age>5</cw-max-age>
<max-cache-wait>20%</max-cache-wait>
<local-port>16720</local-port>
<local-host>http://xxx.xxx.xxx.xxx</local-host>
<hide-names>true</hide-names>
<sync-period>30m</sync-period>
<tracker-url>file:///usr/local/csp/cache/cluster</tracker-url>
<tracker-update>1</tracker-update>
<auto-add-peers>true</auto-add-peers>
<cw-validation checksum="true" zero-counting="true" log-warnings="false" require-id="true"/>
</cache-config>
</cache-handler>

thanks

Display More

Hi.

Did you manage to figure out what was causing this problem, I seem to be suffering from the same thing.

When I start CSP the cache works great with an average of 60-70% cache hits, mostly remote ('R' Flag)

The stats look somthing like below :

avg-received-bytes/s: 0
avg-sent-bytes/s: 7952
cache-peers: 17
cached-ecms: 507
instant-hits: 18745
pending-ecms: 53 (peak: 85)
received-cached: 528812
received-discarded: 251887
received-pending: 454114
received-resendreqs: 135944
remote-hits: 77271
sent-cached: 78119
sent-pending: 74451
sent-resendreqs: 10522
sent-resends: 28013
timeouts: 5892
tracker-failures: 0
wait-hits: 80779

But then after a few hours the 'Pending-ecms' drops to between 0-5 and the 'R' Flags stop. Once the pending ecm count drops the only cache hits i get on my server are local 'C' Flags. Its like it just stops receiving any cache. This happens everytime and i cannot figure out why. If anyone has any idea please let me know.

Thanks, OldStoner.

if someone figured out please tell as i am having same issue my cache starts at 78% and after few hours it comes down to around 40% but if i restart my csp it goes back up.
my setup is
debian server which has two cards on oscam feeding the csp on vps but i dont have any peers in csp all my peers are on another vps running the cccam.
cccam server is conneted to csp with nline.
thanks

guys I'm pretty sure it's because some users that have cw-max-wait time lower than your server time so the client send a service lost message and your csp remove it and after restart it works again but would happening again I suggest to do a map-exclude for that user if you check your logs and the format is like:
Example: <user name="usr2" password="secret2" display-name="User two" profiles="otherprofile" max-connections="1" map-exclude="true"/>

normaly in csp clients update the csp mapper by this way and you can block some of users that may cause this problem by map-exclude in this case :)

I think another way is to set congestion-limit between max-cw-wait/2 and max-cw-wait
So that the the proxy do not consider to set it as congested

i will try the map exclude first because i had false on all and will repost back in few hours thanks for the suggestion :)

ok csp has been running little over 4 hours and cache is still at 80% :)

nice that it works:D

all thanks to you :)
15345972009e0c2559949c9.png

hi sajid. i have the same problem with cluster cache. please advise me how you sorted it. thanx in advance

i have all the user map exclude 'False'.
shall i change it to true for the users causing problem.

Help please