You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: docs/perfsonar/deployment-models.md
+8-7Lines changed: 8 additions & 7 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -10,7 +10,7 @@ each focused on specific set of tests. The following deployment options are curr
10
10
11
11
***Bare metal** - preffered option in one of two possible configurations:
12
12
13
-
```text
13
+
```text
14
14
15
15
* Two bare metal servers, one for latency node, one for bandwidth node
16
16
@@ -20,21 +20,22 @@ each focused on specific set of tests. The following deployment options are curr
20
20
21
21
***Virtual Machine** - if bare metal is not available then it is also possible to run perfSONAR on a VM, however there are a set of additional requirements to fulfill:
22
22
23
-
```text
23
+
```text
24
+
24
25
* Full-node VM is strongly preferred, having 2 VMs (latency/bandwidth node) on a single bare metal. Mixing perfSONAR VM(s) with others might have an impact on the measurements and is therefore not recommended.
25
26
* VM needs to be configured to have SR-IOV to NIC(s) as well as pinned CPUs to ensure bandwidth tests are not impacted (by hypervisor switching CPUs during the test)
26
27
* Succesfull full speed local bandwidth test is highly recommended prior to putting the VM into production
27
28
```
28
29
29
30
***Container** - perfSONAR has supported containers from version 4.1 (Q1 2018) and is documented at <https://docs.perfsonar.net/install_docker.html> but is not typically used in the same way as a full toolkit installation.
30
31
31
-
```text
32
+
```text
32
33
33
34
* Docker perfSONAR test instance can however still be used by sites that run multiple perfSONAR instances on site for their internal testing as this deployment model allows to flexibly deploy a testpoint which can send results to a local measurement archive running on the perfSONAR toolkit node.
34
35
35
36
```
36
37
37
-
###perfSONAR Toolkit vs Testpoint
38
+
## perfSONAR Toolkit vs Testpoint
38
39
39
40
The perfSONAR team has documented the types of installations supported at
40
41
<https://docs.perfsonar.net/install_options.html>. With the release of version 5, OSG/WLCG sites have a new option:
@@ -45,7 +46,7 @@ instead of installing the full Toolkit sites can choose to install the Testpoint
45
46
```text
46
47
* Simpler deployment when a local web interface is not needed and a central measurement archive is available.
47
48
* Less resource intensive for both memory and I/O capacity.
48
-
```
49
+
```
49
50
50
51
* Cons
51
52
@@ -54,7 +55,7 @@ instead of installing the full Toolkit sites can choose to install the Testpoint
54
55
* No web interface to use for configuration or adding local tests
55
56
* Unable to show results in MaDDash
56
57
57
-
```
58
+
```
58
59
59
60
While sites are free to choose whatever deployment method they want, we would like to strongly recommend the use of
60
61
perfSONAR's containerized testpoint. This method was chosen as a "best practice" recommendation because of the reduced
@@ -83,7 +84,7 @@ taking into account the amount of testing that we perform, we recommend at least
83
84
84
85
* NVMe or SSD disk (128GB should be sufficient) if using full Toolkit install with Opensearch.
Copy file name to clipboardExpand all lines: docs/perfsonar/installation.md
+23-11Lines changed: 23 additions & 11 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -48,10 +48,13 @@ You can see more details about EL supported installs at <<https://docs.perfsonar
48
48
49
49
!!! note
50
50
51
-
```text
51
+
<!-- markdownlint-disable MD040 -->
52
+
<!-- markdownlint-disable MD040 -->
53
+
```text
52
54
In all cases, we strongly recommend keeping auto-updates enabled. With yum auto-updates there is a possibility that updated packages can "break" your perfSONAR install but this risk is accepted in order to have security updates quickly applied.
53
55
54
-
```
56
+
```
57
+
<!-- markdownlint-enable MD040 -->
55
58
56
59
The following *additional* steps are needed to configure the toolkit to be used in OSG/WLCG in addition to the steps
57
60
described in the official guide:
@@ -62,7 +65,8 @@ described in the official guide:
62
65
63
66
* You will need to configure your instance(s) to use the OSG/WLCG mesh-configuration. Please follow the steps below:
64
67
65
-
```text
68
+
```text
69
+
66
70
* For toolkit versions 5.0 and higher run: `psconfig remote add https://psconfig.opensciencegrid.org/pub/auto/<FQDN>` replacing `<FQDN>` with your host (e.g. `psum01.aglt2.org`). Verify with `psconfig remote list`.
67
71
```
68
72
@@ -74,15 +78,21 @@ described in the official guide:
74
78
"configure-archives" : true
75
79
}
76
80
]
77
-
```
81
+
```
78
82
79
83
* Please remove any old/stale URLs using `psconfig remote delete <URL>`
80
84
81
-
* If this is a **new instance** or you have changed the node's FQDN, you will need to notify `wlcg-perfsonar-support 'at' cern.ch` to add/update the hostname in one or more test meshes, which will then auto-configure the tests. Please indicate if you have preferences for which meshes your node should be included in (USATLAS, USCMS, ATLAS, CMS, LHCb, Alice, BelleII, etc.). You could also add any additional local tests via web interface (see [Configuring regular tests](http://docs.perfsonar.net/manage_regular_tests.html) for details). Please check which tests are auto-added via central meshes before adding any custom tests to avoid duplication.
85
+
* If this is a **new instance** or you have changed the node's FQDN, you will need to notify
86
+
`wlcg-perfsonar-support 'at' cern.ch` to add/update the hostname in one or more test meshes, which
87
+
will then auto-configure the tests. Please indicate if you have preferences for which meshes your
88
+
node should be included in (USATLAS, USCMS, ATLAS, CMS, LHCb, Alice, BelleII, etc.). You could also
89
+
add any additional local tests via web interface (see [Configuring regular tests](http://docs.perfsonar.net/manage_regular_tests.html) for details).
90
+
Please check which tests are auto-added via central meshes before adding any custom tests to avoid
91
+
duplication.
82
92
83
93
!!! note
84
94
85
-
```text
95
+
```text
86
96
Until your host is added on https://psconfig.opensciencegrid.org to one or more meshes by an administrator the automesh configuration above will not return any tests.
87
97
88
98
```
@@ -127,7 +137,7 @@ site or host firewalls. An overview of perfSONAR security is available at
127
137
```text
128
138
All perfSONAR instances must have port 443 accessible to other perfSONAR instances. Port 443 is used by pScheduler to schedule tests. If unreachable, tests may not run and results may be missing.
129
139
130
-
```
140
+
```
131
141
132
142
For sites that are concerned about having port 443 open, there is a possiblity to get a list of hosts to/from which the
133
143
tests will be initiated. However as this list is dynamic, implementing the corresponding firewall rules would need to be
@@ -140,7 +150,8 @@ network administrators to debug network issues.
140
150
```text
141
151
If you have a central/campus firewall verify required port openings in the perfSONAR security documentation.
142
152
143
-
```
153
+
```
154
+
<!-- markdownlint-enable MD040 -->
144
155
145
156
## Enabling SNMP plugins
146
157
@@ -167,6 +178,7 @@ filling the information please follow those simple guidelines:
167
178
168
179
* For each form (service type) fill at least:
169
180
181
+
<!-- markdownlint-disable MD040 -->
170
182
```text
171
183
172
184
* Hosting Site
@@ -175,7 +187,8 @@ filling the information please follow those simple guidelines:
175
187
* Host IP (optional)
176
188
* Description (optional label used in MaDDash; keep short and unique)
177
189
178
-
```
190
+
```
191
+
<!-- markdownlint-enable MD040 -->
179
192
180
193
* Check "N" when asked "Is it a beta service"
181
194
@@ -186,8 +199,7 @@ filling the information please follow those simple guidelines:
186
199
<!-- -->
187
200
188
201
* GOCDB screen shot for creating a Service Endpoint:
189
-
<img src="../../img/Screen_shot_2013-02-19_at_15.26.52.png" alt="GOCDB screen shot for creating a Service Endpoint"
190
-
width="1024">
202
+

## Gateway requirement, inference, and generator warnings
279
279
280
-
- Any NIC with an IPv4 address should have a corresponding IPv4 gateway; likewise for IPv6. If a NIC lacks a gateway, the generator will attempt conservative inference (below). If a device has no IPv4 or IPv6 gateway (e.g., a management-only NIC), the generator will intentionally skip that NIC when creating an _auto-generated_ config to avoid generating unusable NetworkManager profiles unless you explicitly set the device as `DEFAULT_ROUTE_NIC`.
280
+
- Any NIC with an IPv4 address should have a corresponding IPv4 gateway; likewise for IPv6. If a NIC lacks a
281
+
gateway, the generator will attempt conservative inference (below). If a device has no IPv4 or IPv6 gateway
282
+
(e.g., a management-only NIC), the generator will intentionally skip that NIC when creating an
283
+
_auto-generated_ config to avoid generating unusable NetworkManager profiles unless you explicitly set the
284
+
device as `DEFAULT_ROUTE_NIC`.
281
285
282
286
- Conservative gateway inference: if a NIC has an address/prefix but no gateway, the tool will try to reuse a gateway from another NIC on the SAME subnet.
283
287
@@ -351,7 +355,7 @@ cd docs/perfsonar
351
355
./tests/run_tests.sh
352
356
```
353
357
354
-
## Notes
358
+
## Notes (script details)
355
359
356
360
- The script requires Bash (uses `local -n` namerefs). Run tests on a system
0 commit comments