Chuntao Hong

@rajatmonga Thanks for the explanation. But is there a schedule as to when tensorboard will support other framework? As @zihaolucky has said, this incedent was easily avoidable, should google tensorboard...

经检查，发现SegmentMap会保存目前所有日志文件的handle，只有snapshot的时候才会做truncate操作缩小SegmentMap。所以如果snapshot间隔时间较长就有可能会触发这个bug，而触发之后会出各种诡异问题最终导致数据损坏。 目前看起来可能的解决方法有几个： 1. 在segment.close()里面把handle释放掉，需要读的时候再打开。缺点是读的时候需要检查，且可能有多线程同步问题，实现比较麻烦 2. 在新建segment的时候检查一下SegmentMap大小，如果超过某个阈值就先做一次snapshot然后truncate log。linux默认进程只有1024个handle，阈值设成800之类的应该足够了 3. 在打开handle的地方都加个检查，当handle打开失败的时候做好处理

退出不remove peer的话，leader会一直打"not connected to..."的log啊，braft会把掉线的机器移出去吗？

@lx75249 @zhangchen-qinyinghua

I agree with the subdirectory approach. As for the update hook, I guess we can hook it, but not sure if we can execute python script inside the hook. Anyway,...

For writing your own algorithm, you can try Minerva ( https://github.com/dmlc/minerva), which is also part of DMLC. We are working to integrate CXXNET with Minerva so in the future we...

For doc, please refer to the wiki at: https://github.com/dmlc/minerva/wiki The detailed API doc will be put on repo soon. On Fri, Apr 10, 2015 at 9:57 PM, Eren Golge [email protected]...

Thanks for the info. @minjie, @yutian, could you check if we can use this? On Thu, Jul 23, 2015 at 2:19 PM, futurely [email protected] wrote: > https://github.com/PointCloudLibrary/pcl/blob/master/cmake/pcl_find_cuda.cmake > https://github.com/Itseez/opencv/blob/master/cmake/OpenCVDetectCUDA.cmake >...

No. Sharding is not supported in the open-source version of TuGraph. The commercial version, which will be publicly available at Q4 of 2022, does support sharding.

I think the English version of the docs should be put in /doc