我有一个最近经历MapR cluster
了10 nodes
一些重组。基本上在一些已修复的节点上有一些丢失的磁盘。在此之后,无法在HDFS path
. 但是,hive 表和数据是完整的,可以从 hive 中看到这些表CLI
。
我试过fsck
运行没有任何问题,然后gfsck
执行了。
以下是 gfsck 的输出:
$ sudo /opt/mapr/bin/gfsck -r cluster=tmcluster1 rwvolume=users
Starting GlobalFsck:
clear-mode = false
debug-mode = false
dbcheck-mode = false
repair-mode = true
assume-yes-mode = false
cluster = tmcluster1
rw-volume-name = users
snapshot-name = null
snapshot-id = 0
user-id = 0
group-id = 0
get volume properties ...
put volume users in global-fsck mode ...
get snapshot list for volume users ...
starting phase one (get containers) for volume users(31225146) ...
got volume containers map
done phase one
starting phase two (get inodes) for volume users(31225146) ...
java.lang.Exception: ContainerGetInodes RPC retry 11
at com.mapr.fs.globalfsck.PhaseTwo$PhaseTwoThread.getInodesOfContainer(PhaseTwo.java:238)
at com.mapr.fs.globalfsck.PhaseTwo$PhaseTwoThread.run(PhaseTwo.java:154)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615)
at java.lang.Thread.run(Thread.java:745)
在蜂巢中,我只能看到表格、结构等。我无法触发任何其他DDL
. 例如,表上的正常计数会产生以下错误:
hive> select count(1) from table1;
2015-08-20 21:10:13,4261 ERROR JniCommon fs/client/fileclient/cc/jni_MapRClient.cc:1306 Thread: 27514 mkdirs failed for /user/mapr/tmp/hive/hive_2015-08-20_21-10-13_254_31142713538709, error 11
FAILED: RuntimeException java.io.IOException: Error: Resource temporarily unavailable(11), file: hive_2015-08-20_21-10-13_254_3114271353870903660-1, user name: mapr, ID: 5000
只有一个用户,'mapr'
,用于对 hive 表执行所有操作。
有人对此有任何想法吗?
提前致谢。