A random switching behavior can only be explained by a fluctuating disk space I'd think. Are you running MR operations on the same disk (i.e. is it part of mapred.local.dir as well)?
On Wed, Jan 23, 2013 at 9:24 PM, Mohit Vadhera <project.linux.p...@gmail.com > wrote: > NN switches randomly into the safemode then I run command to leave > safemode manually. I never got alerts for low disk space on machine level > and i didn't see the space fluctuates GBs into MBs . > > > > > > On Wed, Jan 23, 2013 at 9:10 PM, Harsh J <ha...@cloudera.com> wrote: > >> Mohit, >> >> When do you specifically get the error at the NN? Does your NN >> consistently not start with that error? >> >> Your local disk space availability can certainly fluctuate if you use the >> same disk for MR and other activity which creates temporary files. >> >> >> On Wed, Jan 23, 2013 at 9:01 PM, Mohit Vadhera < >> project.linux.p...@gmail.com> wrote: >> >>> Can somebody answer me on this plz ? >>> >>> >>> On Wed, Jan 23, 2013 at 11:44 AM, Mohit Vadhera < >>> project.linux.p...@gmail.com> wrote: >>> >>>> Thanks Guys, As you said the level is already pretty low i.e 100 MB but >>>> in my case the root fs / has 14 G available. What can be the root cause >>>> then ? >>>> >>>> /dev/mapper/vg_operamast1-lv_root >>>> 50G 33G 14G 71% / >>>> >>>> As per logs. >>>> 2013-01-21 01:22:52,217 WARN >>>> org.apache.hadoop.hdfs.server.namenode.NameNodeResourceChecker: Space >>>> available on volume '/dev/mapper/vg_operamast1-lv_root' is 10653696, which >>>> is below the configured reserved amount 104857600 >>>> >>>> >>>> On Wed, Jan 23, 2013 at 11:13 AM, Harsh J <ha...@cloudera.com> wrote: >>>> >>>>> Hi again, >>>>> >>>>> Yes, you need to add it to hdfs-site.xml and restart the NN. >>>>> >>>>> > Thanks Harsh, Do I need to add parameters in hdfs-site.xml and >>>>> restart service namenode. >>>>> > + public static final String DFS_NAMENODE_DU_RESERVED_KEY = >>>>> "dfs.namenode.resource.du. >>>>> reserved"; >>>>> > + public static final long DFS_NAMENODE_DU_RESERVED_DEFAULT = >>>>> 1024 * 1024 * 100; // 100 MB >>>>> >>>>> >>>>> On Wed, Jan 23, 2013 at 10:12 AM, Harsh J <ha...@cloudera.com> wrote: >>>>> >>>>>> Edit your hdfs-site.xml (or whatever place of config your NN uses) to >>>>>> lower the value of property "dfs.namenode.resource.du.reserved". Create a >>>>>> new property if one does not exist, and set the value of space to a >>>>>> suitable level. The default itself is pretty low - 100 MB in bytes. >>>>>> >>>>>> >>>>>> On Wed, Jan 23, 2013 at 9:13 AM, Mohit Vadhera < >>>>>> project.linux.p...@gmail.com> wrote: >>>>>> >>>>>>> Ok Steve. I am forwarding my issue again to the list that you said. >>>>>>> The version is >>>>>>> >>>>>>> Hi, >>>>>>> >>>>>>> Namenode switches into safemode when it has low disk space on the >>>>>>> root fs / i have to manually run a command to leave it. Below are log >>>>>>> messages for low space on root / fs. Is there any parameter so that i >>>>>>> can >>>>>>> reduce reserved amount.Hadoop 2.0.0-cdh4.1.2 >>>>>>> >>>>>>> 2013-01-21 01:22:52,217 WARN >>>>>>> org.apache.hadoop.hdfs.server.namenode.NameNodeResourceChecker: Space >>>>>>> available on volume '/dev/mapper/vg_lv_root' is 10653696, which is below >>>>>>> the configured reserved amount 104857600 >>>>>>> 2013-01-21 01:22:52,218 WARN >>>>>>> org.apache.hadoop.hdfs.server.namenode.FSNamesystem: NameNode low on >>>>>>> available disk space. Entering safe mode. >>>>>>> 2013-01-21 01:22:52,218 INFO org.apache.hadoop.hdfs.StateChange: >>>>>>> STATE* Safe mode is ON. >>>>>>> >>>>>>> >>>>>>> >>>>>>> On Wed, Jan 23, 2013 at 2:50 AM, Steve Loughran < >>>>>>> steve.lough...@gmail.com> wrote: >>>>>>> >>>>>>>> ser@hadoop.apache.orglist >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>> >>>>>> >>>>>> -- >>>>>> Harsh J >>>>>> >>>>> >>>>> >>>>> >>>>> -- >>>>> Harsh J >>>>> >>>> >>>> >>> >> >> >> -- >> Harsh J >> > > -- Harsh J