You are viewing a plain text version of this content. The canonical link for it is here.
Posted to mapreduce-user@hadoop.apache.org by Xizhen Du <xi...@gmail.com> on 2015/11/12 06:35:09 UTC

Log "Start loading edits file" on standby node

Hi All,

We have a standard active/standby hdfs setup, and everything looks working
well by now.

The question is that, on the standby namenode, we are seeing in most of
cases it fetches edits log from 2 journal nodes; but it also happens that
it loads edits from all the 3 journal nodes sometimes, and we don't see
there's a rule-like based on the logging time.

So what's the difference between the 2 behaviors?


# cat /var/log/hadoop-hdfs/hadoop-hdfs-namenode-$HOSTNAME.log | grep -i
"Start loading edits"


++++   from 3 journal nodes:  jn03, jn01, jn02

2015-11-12 12:01:43,296 INFO
org.apache.hadoop.hdfs.server.namenode.FSImage: Start loading edits file
http://jn03:8480/getJournal?jid=thewalking&segmentTxId=43974&storageInfo=-55%3A1852321099%3A0%3ACID-dd00c4ee-90c6-4ff0-9157-59d6645d0ed7,
http://jn01:8480/getJournal?jid=thewalking&segmentTxId=43974&storageInfo=-55%3A1852321099%3A0%3ACID-dd00c4ee-90c6-4ff0-9157-59d6645d0ed7,
http://jn02:8480/getJournal?jid=thewalking&segmentTxId=43974&storageInfo=-55%3A1852321099%3A0%3ACID-dd00c4ee-90c6-4ff0-9157-59d6645d0ed7

++++   from 2 journal nodes: jn01, jn03

2015-11-12 12:03:44,512 INFO
org.apache.hadoop.hdfs.server.namenode.FSImage: Start loading edits file
http://jn01:8480/getJournal?jid=thewalking&segmentTxId=44088&storageInfo=-55%3A1852321099%3A0%3ACID-dd00c4ee-90c6-4ff0-9157-59d6645d0ed7,
http://jn02:8480/getJournal?jid=thewalking&segmentTxId=44088&storageInfo=-55%3A1852321099%3A0%3ACID-dd00c4ee-90c6-4ff0-9157-59d6645d0ed7



Thanks,
Du