-
Notifications
You must be signed in to change notification settings - Fork 3
/
README
executable file
·89 lines (61 loc) · 2.55 KB
/
README
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
QMEM
version: 0.2
release date: 12/12/2012
Author: Txema Heredia Genestar
Email: [email protected]
Institution: Evolutionary Biology Institute (UPF-CSIC)
=========================================================
ABOUT QMEM:
Qmem is a script designed to describe the memory usage of a
SGE cluster. If your cluster has memory restrictions, the
usage of qstat solely is not enough to monitor its state
properly. Qmem attempts to solve that.
=========================================================
INSTALL:
In order to use qmem, your execution hosts must have
configured "PER HOST" the following COMPLEX ATTIBUTES:
slots = set to the host num of processors.
h_vmem = set to the host total memory. Set to
CONSUMABLE
local_disk = set to the host $TMPDIR partition size.
Set to CONSUMABLE
This is the default behaviour. If you want to use your
custom variables, just change the names in the configuration
section at the begining of the script. The slots and memory
variables are mandatory to run qmem. The local_disk one is
optional and you can run qmem without it.
Once you have configured these complex attributes in all
hosts, you are ready to run qmem.
=========================================================
HOW TO USE IT:
qmem Displays the memory usage of the cluster, one line
per host
qmem -u Displays the memory usage of the cluster, node by
node and job by job. Lists all hosts and all jobs
submitted to each host.
-d Shows the real local disk usage on all the cluster
nodes and jobs.
=========================================================
CHANGELOG:
v0.2
- Work by default with slots variable instead of
num_jobs.
- Print host information (global and -u user
information) one by one as they are generated instead
of all at once at the end.
- Read info of new consumable variable local_disk.
- Added new parameter (-d) to report job and node
real disk usage (df/du).
- Fixed host name sorting conflicts with Rocks-Cluster
(Thanks to Angel Carreno).
- Added counter of running & suspended jobs at the
end of output (Thanks to Angel Carreno).
- Fixed color coding escape problems in some terminals
(Thanks to Pablo Escobar for finding it and Jaime
Huerta for fixing it).
=========================================================
ACKNOWLEDGEMENTS:
Pablo Escobar: Find color-coding bug.
Jaime Huerta: Fix color-coding bug.
Angel Carreno: Fix proper sort of qhost.
Add suspended jobs counter