Here is related code from ServerManager#waitForRegionServers() :
int minToStart = this.master.getConfiguration().
getInt(WAIT_ON_REGIONSERVERS_MINTOSTART, defaultMinToStart);
if (minToStart < 1) {
LOG.warn(String.format(
"The value of '%s' (%d) can not be less than 1, ignoring.",
WAIT_ON_REGIONSERVERS_MINTOSTART, minToStart));
minToStart = 1;
}
'hbase.master.wait.on.regionservers.mintostart' is respected.
hbase.master.wait.on.regionservers.interval is used as well.
FYI
On Sat, Sep 2, 2017 at 5:34 AM, Rob V <[email protected]> wrote:
> I need to check if the cloudera restart scripts do this on hbase service
> level. I know for sure they do it when you restart an individual region
> server.
>
> > On 2 Sep 2017, at 12:55, Ted Yu <[email protected]> wrote:
> >
> > Did you use graceful_stop.sh before the restart ?
> >
> >> On Sat, Sep 2, 2017 at 1:12 AM, Rob Verkuylen <[email protected]>
> wrote:
> >>
> >> On CDH5.12 with HBase 1.2, I'm experiencing an issue I thought was long
> >> solved. The regions are all assigned to a single regionserver on a
> restart
> >> of hbase though cloudera manager.
> >>
> >> The Hbase book 9,1.1 states that
> >> 'hbase.master.wait.on.regionservers.mintostart'
> >> should be the proper setting to handle this, but looking at
> ServerManager$
> >> waitForRegionServers this is not a setting that is respected it seems. '
> >> hbase.master.wait.on.regionservers.interval' does seem to be a relevant
> >> setting.
> >>
> >> Am I running into HBASE-14531? Like Ted Yu mentioned in a related post?
> >>
> >> Thanks, Rob
> >>
>
