Circuitbreaker #7532
Circuitbreaker #7532
Conversation
|
Can one of the admins verify this patch? |
14 similar comments
|
Can one of the admins verify this patch? |
|
Can one of the admins verify this patch? |
|
Can one of the admins verify this patch? |
|
Can one of the admins verify this patch? |
|
Can one of the admins verify this patch? |
|
Can one of the admins verify this patch? |
|
Can one of the admins verify this patch? |
|
Can one of the admins verify this patch? |
|
Can one of the admins verify this patch? |
|
Can one of the admins verify this patch? |
|
Can one of the admins verify this patch? |
|
Can one of the admins verify this patch? |
|
Can one of the admins verify this patch? |
|
Can one of the admins verify this patch? |
modules/benchmarks/.factorypath
Outdated
| @@ -0,0 +1,176 @@ | |||
| <factorypath> | |||
There was a problem hiding this comment.
this change of course should be excluded :)
| int lastTrustScore = poolEntry.getTrustScore(); | ||
| long lastSerailId = poolEntry.getSerialId(); | ||
|
|
||
| if (msgSerialId == lastSerailId) { // Pool has rebooted |
There was a problem hiding this comment.
The comment (or logic) is wrong. If the last and new timestamps equal, then the pool is not rebooted
| if (msgSerialId == lastSerailId) { // Pool has rebooted | ||
| nextTrustScore = lastTrustScore/tsDecrease; | ||
| if (nextTrustScore < tsThreshold && !poolEntry.getEnabledStatus()) { // Pool was disabled, should now be re-ENABLED | ||
| LOGGER.error("Pool {} WOULD now be re-ENABLED due to low trust score, BUT IS NOT", poolName); |
There was a problem hiding this comment.
message changed in a5bd221
| } | ||
|
|
||
| public synchronized void messageArrived(CellMessage envelope, PoolManagerPoolUpMessage msg) { | ||
| int tsIncrease = 15; |
There was a problem hiding this comment.
please add comments why this constants are selected
| } | ||
|
|
||
| public synchronized void messageArrived(CellMessage envelope, PoolManagerPoolUpMessage msg) { | ||
| // CostModuleTest#testPoolCircuitbreaker depends on these vaules beeing as they are. |
There was a problem hiding this comment.
Obviously, we don't want test related constants to be a part of the main codebase.
Motivation: The META-INF/gplazma-plugins.xml must point to the correct class name, otherwise we get: Unable register new plugin: Class org.dcache.gplazma.plugins.tokenx.TokenExchange could not be found. Modification: Fix the class name in gplazma-plugins.xml Result: no error on start (and working plugin :) ) Acked-by: Target: master Require-book: no Require-notes: no
…into circuitbreaker
Added Circuitbreaker w/ only logging