Apache
/
hbase
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

HBASE-25594 graceful_stop.sh fails to unload regions when ran at localhost 

Signed-off-by: stack <stack@apache.org>
This commit is contained in:
Javier 2021-02-22 20:51:57 +09:00 committed by stack
parent c36e40eb1c
commit 59ec375d67
2 changed files with 13 additions and 13 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
bin/graceful_stop.sh
View File

@ -103,7 +103,7 @@ hostname=$1
filename="/tmp/$hostname" filename="/tmp/$hostname"
local= local=
localhostname=`/bin/hostname` localhostname=`/bin/hostname -f`
if [ "$localhostname" == "$hostname" ]; then if [ "$localhostname" == "$hostname" ]; then
local=true local=true
@ -201,4 +201,4 @@ else
fi fi
# Cleanup tmp files. # Cleanup tmp files.
trap "rm -f "/tmp/$(basename $0).*.tmp" &> /dev/null" EXIT trap "rm -f /tmp/$(basename $0).*.tmp &> /dev/null" EXIT

22
src/main/asciidoc/_chapters/ops_mgt.adoc
View File

@ -1364,10 +1364,9 @@ Copy the script if you need to make use of it in a version of hbase previous to
A downside to the above stop of a RegionServer is that regions could be offline for a good period of time. A downside to the above stop of a RegionServer is that regions could be offline for a good period of time.
Regions are closed in order. Regions are closed in order.
If many regions on the server, the first region to close may not be back online until all regions close and after the master notices the RegionServer's znode gone. If many regions on the server, the first region to close may not be back online until all regions close and
In Apache HBase 0.90.2, we added facility for having a node gradually shed its load and then shutdown itself down. after the master notices the RegionServer's znode gone. A node can be asked to gradually shed its load and
Apache HBase 0.90.2 added the _graceful_stop.sh_ script. then shutdown itself using the _graceful_stop.sh_ script. Here is its usage:
Here is its usage:
---- ----
$ ./bin/graceful_stop.sh $ ./bin/graceful_stop.sh
@ -1393,16 +1392,17 @@ To decommission a loaded RegionServer, run the following: +$
[NOTE] [NOTE]
==== ====
The `HOSTNAME` passed to _graceful_stop.sh_ must match the hostname that hbase is using to identify RegionServers. The `HOSTNAME` passed to _graceful_stop.sh_ must match the hostname that hbase is using to identify RegionServers.
Check the list of RegionServers in the master UI for how HBase is referring to servers. HBase uses fully-qualified domain names usually. Check the list of RegionServers in the master UI for how HBase
It's usually hostname but can also be FQDN. is referring to servers. Whatever HBase is using, this is what you should pass the _graceful_stop.sh_ decommission script.
Whatever HBase is using, this is what you should pass the _graceful_stop.sh_ decommission script. If you pass IPs, the script is not yet smart enough to make a hostname (or FQDN) of it and so it will fail when it checks
If you pass IPs, the script is not yet smart enough to make a hostname (or FQDN) of it and so it will fail when it checks if server is currently running; the graceful unloading of regions will not run. if server is currently running; the graceful unloading of regions will not run.
==== ====
The _graceful_stop.sh_ script will move the regions off the decommissioned RegionServer one at a time to minimize region churn. The _graceful_stop.sh_ script will move the regions off the decommissioned RegionServer one at a time to minimize region churn.
It will verify the region deployed in the new location before it will moves the next region and so on until the decommissioned server is carrying zero regions. It will verify the region deployed in the new location before it will moves the next region and so on until the decommissioned
At this point, the _graceful_stop.sh_ tells the RegionServer `stop`. server is carrying zero regions. At this point, the _graceful_stop.sh_ tells the RegionServer `stop`.
The master will at this point notice the RegionServer gone but all regions will have already been redeployed and because the RegionServer went down cleanly, there will be no WAL logs to split. The master will at this point notice the RegionServer gone but all regions will have already been redeployed and because the
RegionServer went down cleanly, there will be no WAL logs to split.
[[lb]] [[lb]]
.Load Balancer .Load Balancer