3.1版本存算分离模型创建表超时失败

[托马斯_X]

starrocks 3.1
cdp 7.1.7
单FE和单BE，搭好存算分离模型后尝试建表直接超时失败（默认和60秒都超时）
hdfs里能看到生成的临时文件

求解？

设置超时时间：

be.out (9.2 KB) be.WARNING.log.20230810-132100 (4.2 KB)
be.INFO.log.20230810-131523 (53.5 KB)

[xiangguang]

看 be.out 里有如下错误，应该是 HDFS 集群有问题，先排查下 HDFS 集群的问题吧：

hdfsHFlush: FSDataOutputStream#hflush error:
IOException: Could not get block locations. Source file “/user/root/starrocks/5572960c-bd3e-4c90-8225-b9220e398a9d/10013/SCHEMA_000000000000271E.633a519a-4cb4-4904-b8f0-8a219b026af0.TEMP.” - Aborting…block==nulljava.io.IOException: Could not get block locations. Source file “/user/root/starrocks/5572960c-bd3e-4c90-8225-b9220e398a9d/10013/SCHEMA_000000000000271E.633a519a-4cb4-4904-b8f0-8a219b026af0.TEMP.” - Aborting…block==null
at org.apache.hadoop.hdfs.DataStreamer.setupPipelineForAppendOrRecovery(DataStreamer.java:1525)
at org.apache.hadoop.hdfs.DataStreamer.processDatanodeOrExternalError(DataStreamer.java:1305)
at org.apache.hadoop.hdfs.DataStreamer.run(DataStreamer.java:668)
FSDataOutputStream#close error:
IOException: Could not get block locations. Source file “/user/root/starrocks/5572960c-bd3e-4c90-8225-b9220e398a9d/10013/SCHEMA_000000000000271E.633a519a-4cb4-4904-b8f0-8a219b026af0.TEMP.” - Aborting…block==nulljava.io.IOException: Could not get block locations. Source file “/user/root/starrocks/5572960c-bd3e-4c90-8225-b9220e398a9d/10013/SCHEMA_000000000000271E.633a519a-4cb4-4904-b8f0-8a219b026af0.TEMP.” - Aborting…block==null
at org.apache.hadoop.hdfs.DataStreamer.setupPipelineForAppendOrRecovery(DataStreamer.java:1525)
at org.apache.hadoop.hdfs.DataStreamer.processDatanodeOrExternalError(DataStreamer.java:1305)
at org.apache.hadoop.hdfs.DataStreamer.run(DataStreamer.java:668)

[托马斯_X]

如何排查？目前cm上看整个hadoop集群是正常的，也可以通过hadoop fs -mkdir /user/root/starrocks123 创建目录

[托马斯_X]

防火墙问题，除了8020端口，还会去访问hdfs的9866端口，临时申请了一个同网段虚拟机测通了