[jira] [Updated] (KAFKA-9729) Shrink inWriteLock time in SimpleAuthorizer

["Jiao Zhang (Jira)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/KAFKA-9729?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Jiao Zhang updated KAFKA-9729:
------------------------------
    Description: 
Current SimpleAuthorizer needs 'inWriteLock' when processing add/remove acls requests, while getAcls in authorize() needs 'inReadLock'.
 That means handling add/remove acls requests would block all other requests for example produce and fetch requests.
 When processing add/remove acls, updateResourceAcls() access zk to update acls, which could be long in the case like network glitch.
 We did the simulation for zk delay.
 When adding 100ms delay on zk side, 'inWriteLock' in addAcls()/removeAcls lasts for 400ms~500ms.
 When adding 500ms delay on zk side, 'inWriteLock' in addAcls()/removeAcls lasts for 2000ms~2500ms.
{code:java}
override def removeAcls(aclsTobeRemoved: Set[Acl], resource: Resource): Boolean = {
  inWriteLock(lock) {
    val startMs = Time.SYSTEM.milliseconds()
    val result = updateResourceAcls(resource) { currentAcls =>
      currentAcls -- aclsTobeRemoved
    }
    warn(s"inWriteLock in removeAcls consumes ${Time.SYSTEM.milliseconds() - startMs} milliseconds.")
    result
  }
}{code}
Blocking produce/fetch requests for 2s would cause apparent performance degradation for the whole cluster.
 So considering is it possible to only put 'inWriteLock' inside updateCache. 
{code:java}
// code placeholder
private def updateCache(resource: Resource, versionedAcls: VersionedAcls) {
 if (versionedAcls.acls.nonEmpty) {
 aclCache.put(resource, versionedAcls)
 } else {
 aclCache.remove(resource)
 }
 }
{code}
If do this, block time is only the time for updating local cache, which will not be influenced by network glitch. But don't know if there were special concerns to have current strict write lock and not sure if there are side effects if only put lock to updateCache.

  was:
Current SimpleAuthorizer needs 'inWriteLock' when processing add/remove acls requests, while getAcls in authorize() needs 'inReadLock'.
That means handling add/remove acls requests would block all other requests for example produce and fetch requests.
When processing add/remove acls, updateResourceAcls() access zk to update acls, which could be long in the case like network glitch.
We did the simulation for zk delay.
When adding 100ms delay on zk side, 'inWriteLock' in addAcls()/removeAcls lasts for 400ms~500ms.
When adding 500ms delay on zk side, 'inWriteLock' in addAcls()/removeAcls lasts for 2000ms~2500ms.

Blocking produce/fetch requests for 2s would cause apparent performance degradation for the whole cluster.
So considering is it possible to only put 'inWriteLock' inside updateCache. 
{code:java}
// code placeholder
private def updateCache(resource: Resource, versionedAcls: VersionedAcls) {
 if (versionedAcls.acls.nonEmpty) {
 aclCache.put(resource, versionedAcls)
 } else {
 aclCache.remove(resource)
 }
 }
{code}
If do this, block time is only the time for updating local cache, which will not be influenced by network glitch. But don't know if there were special concerns to have current strict write lock and not sure if there are side effects if only put lock to updateCache.


> Shrink inWriteLock time in SimpleAuthorizer
> -------------------------------------------
>
>                 Key: KAFKA-9729
>                 URL: https://issues.apache.org/jira/browse/KAFKA-9729
>             Project: Kafka
>          Issue Type: Improvement
>          Components: security
>    Affects Versions: 1.1.0
>            Reporter: Jiao Zhang
>            Priority: Minor
>
> Current SimpleAuthorizer needs 'inWriteLock' when processing add/remove acls requests, while getAcls in authorize() needs 'inReadLock'.
>  That means handling add/remove acls requests would block all other requests for example produce and fetch requests.
>  When processing add/remove acls, updateResourceAcls() access zk to update acls, which could be long in the case like network glitch.
>  We did the simulation for zk delay.
>  When adding 100ms delay on zk side, 'inWriteLock' in addAcls()/removeAcls lasts for 400ms~500ms.
>  When adding 500ms delay on zk side, 'inWriteLock' in addAcls()/removeAcls lasts for 2000ms~2500ms.
> {code:java}
> override def removeAcls(aclsTobeRemoved: Set[Acl], resource: Resource): Boolean = {
>   inWriteLock(lock) {
>     val startMs = Time.SYSTEM.milliseconds()
>     val result = updateResourceAcls(resource) { currentAcls =>
>       currentAcls -- aclsTobeRemoved
>     }
>     warn(s"inWriteLock in removeAcls consumes ${Time.SYSTEM.milliseconds() - startMs} milliseconds.")
>     result
>   }
> }{code}
> Blocking produce/fetch requests for 2s would cause apparent performance degradation for the whole cluster.
>  So considering is it possible to only put 'inWriteLock' inside updateCache. 
> {code:java}
> // code placeholder
> private def updateCache(resource: Resource, versionedAcls: VersionedAcls) {
>  if (versionedAcls.acls.nonEmpty) {
>  aclCache.put(resource, versionedAcls)
>  } else {
>  aclCache.remove(resource)
>  }
>  }
> {code}
> If do this, block time is only the time for updating local cache, which will not be influenced by network glitch. But don't know if there were special concerns to have current strict write lock and not sure if there are side effects if only put lock to updateCache.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)