https://jira.opendaylight.org This file is an XML representation of an issue en-us 8.20.10 820010 22-06-2022 https://jira.opendaylight.org/browse/YANGTOOLS-165 yangtools <p>Durting performance testing we have found that OF plugin keeps a strong-referenced cache, which is only boiund by time and dominated by strong references.</p> <p>This means that as throughput goes up, so will the memory pressure.</p> <p>Make the depth of the cache configurable and make it use softvalues, so that it a) has an upper bound on memory retained irrespective of throughput 2) reacts to memory pressure by sacrificing this information instead of running out of memory.</p> <p>Operating System: Linux<br/> Platform: PC</p> YANGTOOLS-165

OFplugin has a strong-referenced unbounded cache

Bug Verified Done Michal Rehak Robert Varga Wed, 14 May 2014 17:21:22 +0000 Sun, 10 Apr 2022 18:35:06 +0000 Fri, 16 May 2014 15:52:19 +0000 0 2 <p>So the cache is used to support a 'replay' history thing. A cache is the wrong data structure to support this use case, as its performance decreases with elements retained. The time-based eviction policy incurs heavy penalty when retention time is exceeded.</p> <p>What you really want is a fixed-depth ArrayBlockingQueue. Make the depth configurable. Then to store a new packet do:</p> <p>ArrayBlockingQueue q;<br/> Object msg;</p> <p>if (!q.offer(msg)) {<br/> q.poll();<br/> q.offer(msg);<br/> }</p> <p>which will give you a guaranteed replay buffer, with excellent O(1) performance. Given the high number of switches we want to support I would suggest to wrap the message in a SoftReference so we automatically throw debug data out before runnig out of memory.</p> <p>Note that the ArrayBlockingQueue uses a preallocated backing array, so be careful to use reasonable numbers.</p> <p>If you need time-based eviction, you can wrap the message into:</p> <p>class MessageTime {<br/> Object msg;<br/> long time = System.nanoTime();<br/> }</p> <p>before inserting into the queue. Then have a background thread run the following periodically:</p> <p>long cutoff = System.nanoTime() - RETENTION_TIME;<br/> while (true) { // Not really, place an upper bound on items processed!<br/> MessageTime m = q.peek();<br/> if (m == null || m.time &gt;= cutoff) </p> { break; } <p> q.remove(m); // Because an enqueue thread may have removed it already<br/> }</p> <p><a href="https://git.opendaylight.org/gerrit/#/c/6994/" class="external-link" target="_blank" rel="nofollow noopener">https://git.opendaylight.org/gerrit/#/c/6994/</a></p> <p>removing bulk transaction cache<br/> <a href="https://git.opendaylight.org/gerrit/#/c/7024/" class="external-link" target="_blank" rel="nofollow noopener">https://git.opendaylight.org/gerrit/#/c/7024/</a></p> <p>cbench is showing speed improvements (up to 5%), memory consumption is lower (200MB)</p> Development External issue ID 1006 External issue URL Rank 0|i025jj: