2013年5月30日 星期四
2013年5月8日 星期三
[Emulab 管理] Node Status欄位值
狀況：如Node仍無法自己重reloading中跑回free，檢查「Startup Status」欄位。
檢查點：「none」為正常，「1」為有問題
處理方式：update nodes set startstatus='none' where node_id='<Node_id>';
狀況外：
如果仍不可解決，先丟回hwdown
[Emulab 管理] 砍使用者、Node、Project
del user：
1. web 砍帳號
2. delete from users where uid="帳號";
砍Node：
withadminprivs /usr/testbed/sbin/deletenode pc35
----------------------------------------------------------------------------
砍project：
delete from projects where pid='xxx';
delete from users where uid='hmlee';
[Emulab 管理] ErrorNum was -24
boss# wap snmpit -l
*** WARNING: snmpit:
*** SNMP GET failed - In snmpit
*** SNMPIT get failed for device es1 (try 30 of 30)
*** Variable was [sysDescr,0,(undefined),(undefined)]
*** Returned (undefined), ErrorNum was -24
*** Error string is: Timeout
RPC::Async::Server: Disconnecting client for error: SNMP GET failed - In snmpit
SNMPIT get failed for device es1 (try 30 of 30)
Variable was [sysDescr,0,(undefined),(undefined)]
Returned (undefined), ErrorNum was -24
Error string is: Timeout
原因：
snmp_community設錯
處理方式：
mysql> update switch_stack_types set snmp_community='private' where stack_id='Experiment';
[系統] 監控HD容量
檔名：/shell/disk.sh
#####################
說明：利用 mount 指令解決 Read-only file system 問題
#####################
#!/bin/sh
ADMIN="xxx@crypto.ee.ncku.edu.tw"
# set alert level 90% is default
ALERT=90
day=`date +"%Y%m%d%H%M"`
df -H | grep -vE '^Filesystem|dev/da0s1a|devfs' | awk '{ print $5 " " $6 }' | while read output;
do
echo $output
usep=$(echo $output | awk '{ print $1}' | cut -d'%' -f1 )
partition=$(echo $output | awk '{ print $2 }' )
if [ $usep -ge $ALERT ]; then
echo "Running out of space \"$partition ($usep%)\" on $(hostname) as on $(date)" |
mail -s "Alert: Almost out of disk space $usep%" $ADMIN
# clear ops server /var/mail/root file
if [ $partition=="/usr/testbed/usersvar" ]; then
mount -o rw /usr/testbed/usersvar
cp -p /usr/testbed/usersvar/mail/root /mnt/backup2/ops_var_log_bak/mail/root."$day"
cat /dev/null > /usr/testbed/usersvar/mail/root
umount /usr/testbed/usersvar
fi
# clear boss server /var/log
if [ $partition=="/var" ]; then
cp -p /var/log/auth.log /mnt/backup2/boss_var_log_bak/auth.log."$day"
cat /dev/null > /var/log/auth.log
cp -p /var/log/maillog /mnt/backup2/boss_var_log_bak/maillog."$day"
cat /dev/null > /var/log/maillog
fi
fi
done
-----------------------------------------------------------------------
修改crontab
#ee /etc/crontab
-----------------------------------------------------------------------
新增底下二行!! 每5分鐘檢查一次
# mointor disk
*/5 * * * * root /shell/disk.sh
[Emulab 管理] Node卡在reloading處理方式
root@boss[/root]# sched_reserve emulab-ops hwdown <Node_id>
Checking if <Node_id> is reserved...Reserved - Scheduling next reservation...
root@boss[/root]# su adminemu
% nfree emulab-ops reloading <Node_id>
Moving<Node_id> to emulab-ops/hwdown.
% nfree emulab-ops hwdown <Node_id>
Moving
<Node_id>
to emulab-ops/reloaing.
註解：先將Node丟回hwdown，再重新丟Free之(reloading)，Node將重新開機
[Emulab 管理] 手動指派Node到Proj
清除指令：
update nodes set reserved_pid=NULL;
[Emulab 管理] 更換主機版設定
如有更換主機版機器，網卡需重新設定。
於資料庫更新MAC
update interfaces set mac='0090fb2f3b56' where node_id='pc21' and card='0';
update interfaces set mac='0090fb2f3b57' where node_id='pc21' and card='1';
update interfaces set mac='0090fb2f3b58' where node_id='pc21' and card='2';
update interfaces set mac='0090fb2f3b59' where node_id='pc21' and card='3';
update interfaces set mac='0090fb2f3b5a' where node_id='pc21' and card='4';
update interfaces set mac='0090fb2f3b5b' where node_id='pc21' and card='5';
[Emulab 管理] 指派單一node到已存在的實驗內
指令：nalloc [-d] [-p] <pid> <eid> <node> <node> <...>
root@boss[/root]#su adminemu
You have mail.
%wap nalloc cryptolab TWISC-Project pc145
[Emulab 管理] Web介面限制使用者登入指令
限制所有成員之網頁介面登入指令：update sitevariables set value='1' where name='web/nologins';
開發所有成員之網頁介面登入指令：update sitevariables set value='0' where name='web/nologins';
限制特定Project成員之網頁介面登入指令：
update users u inner join group_membership g on u.uid=g.uid set
weblogin_frozen=1
where pid='NS-NCKUEE';
開放特定Project成員之網頁介面登入指令：
update users u inner join group_membership g on u.uid=g.uid set
weblogin_frozen=0
where pid='NS-NCKUEE';
[Emulab 管理] 重載Image指令
大眾用指令：os_load -i <imageid> <node>
Image專屬某Porject指令：os_load [-p <PID>] -i <imageid> <node>
Image專屬某Porject指令：os_load [-p <PID>] -i <imageid> <node>
2013年5月3日 星期五
訂閱：
文章 (Atom)