1 | DESCRIPTION |
---|
2 | =========== |
---|
3 | |
---|
4 | Job Monarch is a set of tools to monitor and optionally archive (batch)job information. |
---|
5 | |
---|
6 | It is a addon for the Ganglia monitoring system and plugs in to a existing Ganglia setup. |
---|
7 | |
---|
8 | To view a operational setup with Job Monarch, have a look here: http://ganglia.sara.nl/ |
---|
9 | |
---|
10 | |
---|
11 | Job Monarch stands for 'Job Monitoring and Archiving' tool and consists of three (3) components: |
---|
12 | |
---|
13 | * jobmond |
---|
14 | |
---|
15 | The Job Monitoring Daemon. |
---|
16 | |
---|
17 | Gathers PBS/Torque batch statistics on jobs/nodes and submits them into |
---|
18 | Ganglia's XML stream. |
---|
19 | |
---|
20 | Through this daemon, users are able to view the PBS/Torque batch system and the |
---|
21 | jobs/nodes that are in it (be it either running or queued). |
---|
22 | |
---|
23 | * jobarchived (optionally) |
---|
24 | |
---|
25 | The Job Archiving Daemon. |
---|
26 | |
---|
27 | Listens to Ganglia's XML stream and archives the job and node statistics. |
---|
28 | It stores the job statistics in a Postgres SQL database and the node statistics |
---|
29 | in RRD files. |
---|
30 | |
---|
31 | Through this daemon, users are able to lookup a old/finished job |
---|
32 | and view all it's statistics. |
---|
33 | |
---|
34 | Optionally: You can either choose to use this daemon if your users have use for it. |
---|
35 | As it can be a heavy application to run and not everyone may have a need for it. |
---|
36 | |
---|
37 | - Multithreaded: Will not miss any data regardless of (slow) storage |
---|
38 | |
---|
39 | - Staged writing: Spread load over bigger time periods |
---|
40 | |
---|
41 | - High precision RRDs: Allow for zooming on old periods with large precision |
---|
42 | |
---|
43 | - Timeperiod RRDs: Allow for smaller number of files while still keeping advantage |
---|
44 | of small disk space |
---|
45 | |
---|
46 | * web |
---|
47 | |
---|
48 | The Job Monarch web interface. |
---|
49 | |
---|
50 | This interfaces with the jobmond data and (optionally) the jobarchived and presents the |
---|
51 | data and graphs. |
---|
52 | |
---|
53 | It does this in a similar layout/setup as Ganglia itself, so the navigation and usage is intuitive. |
---|
54 | |
---|
55 | - Graphical usage: Displays graphical cluster overview so you can see the cluster (job) state |
---|
56 | in one view/image and additional pie chart with relevant information on your |
---|
57 | current view |
---|
58 | |
---|
59 | - Filters: Ability to filter output to limit information displayed (usefull for those |
---|
60 | clusters with 500+ jobs). This also filters the graphical overview images output |
---|
61 | and pie chart so you only see the filter relevant data |
---|
62 | |
---|
63 | - Archive: When enabling jobarchived, users can go back as far as recorded in the database |
---|
64 | or archived RRDs to find out what happened to a crashed or old job |
---|
65 | |
---|
66 | - Zoom ability: Users can zoom into a timepriod as small as the smallest grain of the RRDS |
---|
67 | (typically up to 10 seconds) when a jobarchived is present |
---|
68 | |
---|
69 | REQUIREMENTS |
---|
70 | ============ |
---|
71 | |
---|
72 | all: |
---|
73 | |
---|
74 | - Python 2.3 or higher |
---|
75 | |
---|
76 | jobmond: |
---|
77 | |
---|
78 | - pbs_python v2.8.2 or higher |
---|
79 | ftp://ftp.sara.nl/pub/outgoing/pbs_python.tar.gz |
---|
80 | |
---|
81 | - gmond v3.0.1 or higher |
---|
82 | http://www.ganglia.info |
---|
83 | |
---|
84 | jobarchived: |
---|
85 | |
---|
86 | - Postgres SQL v7.xx |
---|
87 | http://www.postgres.org |
---|
88 | |
---|
89 | - rrdtool v1.xx |
---|
90 | http://people.ee.ethz.ch/~oetiker/webtools/rrdtool/ |
---|
91 | |
---|
92 | - python-pgsql v4.x.x |
---|
93 | http://sourceforge.net/projects/pypgsql/ |
---|
94 | |
---|
95 | - gmetad v3.x.x |
---|
96 | http://www.ganglia.info |
---|
97 | |
---|
98 | web: |
---|
99 | |
---|
100 | - PHP v4.1 or higher |
---|
101 | http://www.php.net |
---|
102 | |
---|
103 | - php-pgsql v4.x.x |
---|
104 | (should come with Postgres) |
---|
105 | |
---|
106 | - GD v2.x |
---|
107 | http://www.boutell.com/gd/ |
---|
108 | |
---|
109 | - Ganglia web frontend v3.x.x |
---|
110 | http://www.ganglia.info |
---|
111 | |
---|
112 | |
---|
113 | INSTALLATION |
---|
114 | ============ |
---|
115 | |
---|
116 | Prior to installing the software make sure you meet the necessary requirements as |
---|
117 | mentioned above. |
---|
118 | |
---|
119 | NOTE: You can choose to install to other path/directories if your setup is different. |
---|
120 | |
---|
121 | * jobmond |
---|
122 | |
---|
123 | 1. Copy jobmond.py: |
---|
124 | |
---|
125 | > cp jobmond/jobmond.py /usr/local/sbin/jobmond.py |
---|
126 | |
---|
127 | 2. Copy jobmond.conf: |
---|
128 | |
---|
129 | > cp jobmond/jobmond.conf /etc/jobmond.conf |
---|
130 | |
---|
131 | * jobarchived |
---|
132 | |
---|
133 | 1. Create a Postgres SQL database for jobarchived: |
---|
134 | |
---|
135 | > createdb jobarchive |
---|
136 | |
---|
137 | 2. Setup jobarchived's tables: |
---|
138 | |
---|
139 | > psql -f jobarchived/job_dbase.sql jobarchive |
---|
140 | |
---|
141 | 3. Copy jobarchived/jobarchived.conf: |
---|
142 | |
---|
143 | > cp jobarchived/jobarchived.conf /etc/jobarchived.conf |
---|
144 | |
---|
145 | 4. Copy jobarchived.py and DBClass.py: |
---|
146 | |
---|
147 | > cp jobarchived/jobarchived.py /usr/local/sbin/jobarchived.py |
---|
148 | > cp jobarchived/DBClass.py /usr/local/sbin/DBClass.py |
---|
149 | |
---|
150 | * web |
---|
151 | |
---|
152 | 1. Copy the Job Monarch Template to your Ganglia installation |
---|
153 | |
---|
154 | > cp -a web/templates/job_monarch /var/www/ganglia/templates |
---|
155 | |
---|
156 | 2. Copy the web interface files to the addon directory in Ganglia |
---|
157 | |
---|
158 | > cp -a web/addons/job_monarch /var/www/ganglia/addons |
---|
159 | |
---|
160 | CONFIGURATION |
---|
161 | ============= |
---|
162 | |
---|
163 | After installation each component requires additional configuration. |
---|
164 | |
---|
165 | * jobmond |
---|
166 | |
---|
167 | 1. Edit Jobmond's config to reflect your settings: |
---|
168 | |
---|
169 | - In /etc/jobmond.conf |
---|
170 | |
---|
171 | ( see config comments for syntax and explanation ) |
---|
172 | |
---|
173 | * jobarchived |
---|
174 | |
---|
175 | 1. Edit Jobarchived's config to reflect your settings: |
---|
176 | |
---|
177 | - In /etc/jobarchived.conf |
---|
178 | |
---|
179 | ( see config comments for syntax and explanation ) |
---|
180 | |
---|
181 | * web |
---|
182 | |
---|
183 | 1. Change your Ganglia's web template to Job Monarch |
---|
184 | |
---|
185 | - In /var/www/ganglia/conf.php: |
---|
186 | |
---|
187 | > $template_name = "job_monarch"; |
---|
188 | |
---|
189 | 2. Change Job Monarch's config to reflect your settings: |
---|
190 | |
---|
191 | - In /var/www/ganglia/addons/job_monarch/conf.php |
---|
192 | |
---|
193 | ( see config comments for syntax and explanation ) |
---|
194 | |
---|
195 | START |
---|
196 | ===== |
---|
197 | |
---|
198 | * jobmond |
---|
199 | |
---|
200 | The Job Monitor has to be run on a machine that is allowed to |
---|
201 | query the PBS/Torque server. |
---|
202 | Make sure that if you have 'acl_hosts' enabled on your PBS/Torque |
---|
203 | server that jobmond's machine is in it. |
---|
204 | |
---|
205 | 1. Start the Job Monitor: |
---|
206 | |
---|
207 | > /usr/local/sbin/jobmond.py -c /etc/jobmond.conf |
---|
208 | |
---|
209 | * jobarchived |
---|
210 | |
---|
211 | 1. Start the Job Archiver: |
---|
212 | |
---|
213 | > /usr/local/sbin/jobarchived.py -c /etc/jobarchived.conf |
---|
214 | |
---|
215 | * web |
---|
216 | |
---|
217 | Doesn't require you to (re)start anything. |
---|
218 | ( make sure the Postgres database is running though ) |
---|
219 | |
---|
220 | CONTACT |
---|
221 | ======= |
---|
222 | |
---|
223 | To contact the author for anything from bugfixes to flame/hate mail: |
---|
224 | |
---|
225 | * Ramon Bastiaans |
---|
226 | |
---|
227 | <ramon ( a t ) sara ( d o t ) nl> |
---|