diff options
Diffstat (limited to 'book-enea-nfv-core-release-info/doc/known_bugs_and_limitations.xml')
-rw-r--r-- | book-enea-nfv-core-release-info/doc/known_bugs_and_limitations.xml | 356 |
1 files changed, 356 insertions, 0 deletions
diff --git a/book-enea-nfv-core-release-info/doc/known_bugs_and_limitations.xml b/book-enea-nfv-core-release-info/doc/known_bugs_and_limitations.xml new file mode 100644 index 0000000..039fae9 --- /dev/null +++ b/book-enea-nfv-core-release-info/doc/known_bugs_and_limitations.xml | |||
@@ -0,0 +1,356 @@ | |||
1 | <?xml version="1.0" encoding="ISO-8859-1"?> | ||
2 | <!DOCTYPE chapter PUBLIC "-//OASIS//DTD DocBook XML V4.2//EN" | ||
3 | "http://www.oasis-open.org/docbook/xml/4.2/docbookx.dtd"> | ||
4 | <chapter id="bugs-limitations"> | ||
5 | <title>Known Issues and Limitations</title> | ||
6 | |||
7 | <para>Problems and limitations relevant to the current release are detailed | ||
8 | in this chapter. Corrections to bugs detected by Enea are submitted | ||
9 | upstream, and remaining issues are listed below.<remark>INFO: The <emphasis | ||
10 | role="bold">Release-Specific Problems</emphasis> section further down is | ||
11 | generated from JIRA with gen_known_issues.py, but that script is HARDCODED | ||
12 | with affectedversion "Enea NFV 1.0" and needs to be adapted when a release | ||
13 | info for another ENFV version changes.</remark></para> | ||
14 | |||
15 | <section id="bugs-limitations-gen"> | ||
16 | <title>Release Specific Issues</title> | ||
17 | |||
18 | <orderedlist> | ||
19 | <listitem> | ||
20 | <para>Security groups are not working correctly for ICMP traffic in | ||
21 | deployments with OpenDaylight.</para> | ||
22 | |||
23 | <itemizedlist> | ||
24 | <listitem> | ||
25 | <para>Description and Impact:</para> | ||
26 | |||
27 | <orderedlist> | ||
28 | <listitem> | ||
29 | <para>When OPNFV is deployed with OpenDaylight as an SDN | ||
30 | controller, the Security Groups rules pertaining to ICMP do | ||
31 | not work as expected. The OpenFlow rules describing the ICMP | ||
32 | rules are inconsistent, so VMs can be pinged even when this is | ||
33 | not desired.</para> | ||
34 | </listitem> | ||
35 | |||
36 | <listitem> | ||
37 | <para>This reproduces on aarch64. On x86 the security groups | ||
38 | work correctly.</para> | ||
39 | </listitem> | ||
40 | </orderedlist> | ||
41 | </listitem> | ||
42 | |||
43 | <listitem> | ||
44 | <para>Workaround: N/A.</para> | ||
45 | </listitem> | ||
46 | </itemizedlist> | ||
47 | </listitem> | ||
48 | |||
49 | <listitem> | ||
50 | <para>Virtual instances do not get IP from DHCP in SFC scenarios with | ||
51 | ODL</para> | ||
52 | |||
53 | <itemizedlist> | ||
54 | <listitem> | ||
55 | <para>Description and Impact:</para> | ||
56 | |||
57 | <orderedlist> | ||
58 | <listitem> | ||
59 | <para>After a fresh deploy of OPNFV with OpenDaylight and the | ||
60 | Service Function Chaining scenario configurations, instances | ||
61 | fail to get IP from DHCP, due to OpenDaylight | ||
62 | malfunctioning.</para> | ||
63 | </listitem> | ||
64 | |||
65 | <listitem> | ||
66 | <para>The SFC VNFs are not reachable via SSH for management | ||
67 | and configuration.</para> | ||
68 | </listitem> | ||
69 | </orderedlist> | ||
70 | </listitem> | ||
71 | |||
72 | <listitem> | ||
73 | <para>Workaround: Restarting OpenDaylight via | ||
74 | <literal>systemctl</literal> fixes the problem.</para> | ||
75 | </listitem> | ||
76 | </itemizedlist> | ||
77 | </listitem> | ||
78 | |||
79 | <listitem> | ||
80 | <para>Scale-in and Scale-out operations are not supported</para> | ||
81 | |||
82 | <itemizedlist> | ||
83 | <listitem> | ||
84 | <para>Description and Impact:</para> | ||
85 | |||
86 | <orderedlist> | ||
87 | <listitem> | ||
88 | <para>When adding or removing nodes via Fuel installer on an | ||
89 | OPNFV deployment that uses OpenDaylight, flows in | ||
90 | <filename>ovs</filename> tables on the existing and new nodes | ||
91 | are not updated properly. Similarly, OpenvSwitch ports | ||
92 | interfaces are not removed when deleting nodes.</para> | ||
93 | </listitem> | ||
94 | |||
95 | <listitem> | ||
96 | <para>As a consequence, when adding a compute node to an | ||
97 | existing OPNFV deployment, the instances running on the new | ||
98 | compute node cannot be reached by floating IP. Scale-out and | ||
99 | scale-in operations are flawed.</para> | ||
100 | </listitem> | ||
101 | </orderedlist> | ||
102 | </listitem> | ||
103 | |||
104 | <listitem> | ||
105 | <para>Workaround: The rules can be manually configured, but this | ||
106 | is not scalable.</para> | ||
107 | </listitem> | ||
108 | </itemizedlist> | ||
109 | </listitem> | ||
110 | |||
111 | <listitem> | ||
112 | <para>Hot-plugging network interfaces does not work in aarch64 | ||
113 | guests</para> | ||
114 | |||
115 | <itemizedlist> | ||
116 | <listitem> | ||
117 | <para>Description and Impact:</para> | ||
118 | |||
119 | <orderedlist> | ||
120 | <listitem> | ||
121 | <para>When the user tries to hot-plug a network port to a | ||
122 | running instance, this fails due to virtio negotiation failure | ||
123 | between the host and guest. This happens when the VM is using | ||
124 | a Linux Kernel version prior to 4.10. Port addition or | ||
125 | deletion is not possible.</para> | ||
126 | </listitem> | ||
127 | |||
128 | <listitem> | ||
129 | <para>This reproduces on aarch64.</para> | ||
130 | </listitem> | ||
131 | </orderedlist> | ||
132 | </listitem> | ||
133 | |||
134 | <listitem> | ||
135 | <para>Workaround: The user needs to upgrade the guest Linux Kernel | ||
136 | to 4.10 or later. In case of port addition, the user also needs to | ||
137 | manually set the link up and trigger the | ||
138 | <literal>dhclient</literal> request.</para> | ||
139 | </listitem> | ||
140 | </itemizedlist> | ||
141 | </listitem> | ||
142 | |||
143 | <listitem> | ||
144 | <para>Fuel Healthcheck Configuration tests fail when default | ||
145 | credentials are used</para> | ||
146 | |||
147 | <itemizedlist> | ||
148 | <listitem> | ||
149 | <para>Description and Impact:</para> | ||
150 | |||
151 | <orderedlist> | ||
152 | <listitem> | ||
153 | <para>The Configuration tests from the Fuel Healthcheck fail | ||
154 | when the OPNFV cluster operates with the default credentials. | ||
155 | This has no impact on the overall cluster | ||
156 | functionality.</para> | ||
157 | </listitem> | ||
158 | |||
159 | <listitem> | ||
160 | <para>This is architecture independent.</para> | ||
161 | </listitem> | ||
162 | </orderedlist> | ||
163 | </listitem> | ||
164 | |||
165 | <listitem> | ||
166 | <para>Workaround: In order to make these 3 test cases pass, the | ||
167 | user needs to:</para> | ||
168 | |||
169 | <orderedlist> | ||
170 | <listitem> | ||
171 | <para>Change the credentials for the root user on the Fuel | ||
172 | Master node: after logging on as root on the master node, | ||
173 | issue command <command>passwd</command> and provide the new | ||
174 | password.</para> | ||
175 | </listitem> | ||
176 | |||
177 | <listitem> | ||
178 | <para>Change the default credentials for the OpenStack | ||
179 | cluster: in Fuel Dashboard, for an Environment, change | ||
180 | Settings|General|OpenStack Access > Password to something | ||
181 | other than the default.</para> | ||
182 | </listitem> | ||
183 | |||
184 | <listitem> | ||
185 | <para>Change the default credentials for Keystone on the | ||
186 | master node: in Fuel Dashboard, click on the user avatar and | ||
187 | select <literal>Change Password</literal>. This should be done | ||
188 | before deploy or you will need to redeploy to have the changes | ||
189 | applied.</para> | ||
190 | </listitem> | ||
191 | </orderedlist> | ||
192 | </listitem> | ||
193 | </itemizedlist> | ||
194 | </listitem> | ||
195 | |||
196 | <listitem> | ||
197 | <para>Fuel Healthcheck Stack update test fails</para> | ||
198 | |||
199 | <itemizedlist> | ||
200 | <listitem> | ||
201 | <para>Description and Impact:</para> | ||
202 | |||
203 | <para>The Platform test case number 5 (Update stack) from the Fuel | ||
204 | Healthcheck fails. This has no impact on the overal cluster | ||
205 | functionality.</para> | ||
206 | </listitem> | ||
207 | |||
208 | <listitem> | ||
209 | <para>Workaround: N/A.</para> | ||
210 | </listitem> | ||
211 | </itemizedlist> | ||
212 | </listitem> | ||
213 | |||
214 | <listitem> | ||
215 | <para>Issue #1 with Openstack Resource Agents and Compute Fencing | ||
216 | functionality</para> | ||
217 | |||
218 | <itemizedlist> | ||
219 | <listitem> | ||
220 | <para>Description and Impact:</para> | ||
221 | |||
222 | <para>In an OPNFV deployment that uses Openstack Resource Agents, | ||
223 | the neutron-openvswitch-agent is killed by Pacemaker when booting, | ||
224 | due to Pacemaker misconfiguration.</para> | ||
225 | </listitem> | ||
226 | |||
227 | <listitem> | ||
228 | <para>Workaround:</para> | ||
229 | |||
230 | <para>Starting the <literal>systemd</literal> service manually | ||
231 | makes it run successfully. Enea NFV Core 1.0 will be shipped | ||
232 | without Openstack Resource Agents, therefore this issue should not | ||
233 | affect the user.</para> | ||
234 | </listitem> | ||
235 | </itemizedlist> | ||
236 | </listitem> | ||
237 | |||
238 | <listitem> | ||
239 | <para>Virtual instances are not affected by removing a node from Ceph | ||
240 | Storage Cluster</para> | ||
241 | |||
242 | <itemizedlist> | ||
243 | <listitem> | ||
244 | <para>Description and Impact:</para> | ||
245 | |||
246 | <para>Engineering wanted to validate the survival of the storage | ||
247 | systems when a single disk is removed without causing data loss. | ||
248 | Without physical access to the test setup, this test is not | ||
249 | feasible.</para> | ||
250 | </listitem> | ||
251 | |||
252 | <listitem> | ||
253 | <para>Workaround:</para> | ||
254 | |||
255 | <itemizedlist> | ||
256 | <listitem> | ||
257 | <para>The chosen approach was to validate what happens to the | ||
258 | Ceph cluster when network connectivity is lost for the Storage | ||
259 | interface of one of the nodes. The observed behavior: no | ||
260 | impact on a running instance using Ceph for volume | ||
261 | storage.</para> | ||
262 | </listitem> | ||
263 | |||
264 | <listitem> | ||
265 | <para><ulink | ||
266 | url="http://ceph.com/geen-categorie/admin-guide-replacing-a-failed-disk-in-a-ceph-cluster/">Reference | ||
267 | information</ulink></para> | ||
268 | </listitem> | ||
269 | </itemizedlist> | ||
270 | </listitem> | ||
271 | </itemizedlist> | ||
272 | </listitem> | ||
273 | |||
274 | <listitem> | ||
275 | <para>Issue #2 with Openstack Resource Agents and Compute Fencing | ||
276 | functionality</para> | ||
277 | |||
278 | <itemizedlist> | ||
279 | <listitem> | ||
280 | <para>Description and Impact:</para> | ||
281 | |||
282 | <para>In an OPNFV deployment that uses Openstack Resource Agents, | ||
283 | when we configure the <literal>fence_compute</literal> as a | ||
284 | Pacemaker resource, the Controller nodes start to reboot each | ||
285 | other endlessly.</para> | ||
286 | </listitem> | ||
287 | |||
288 | <listitem> | ||
289 | <para>Workaround:</para> | ||
290 | |||
291 | <para>Enea NFV Core 1.0 will be shipped without Openstack Resource | ||
292 | Agents, therefore this issue should not affect the user.</para> | ||
293 | </listitem> | ||
294 | </itemizedlist> | ||
295 | </listitem> | ||
296 | |||
297 | <listitem> | ||
298 | <para>Offline Deploy with Fuel fails at times</para> | ||
299 | |||
300 | <itemizedlist> | ||
301 | <listitem> | ||
302 | <para>Description and Impact:</para> | ||
303 | |||
304 | <para>The Postfix installation and post-install configuration | ||
305 | sometimes take longer when the Fuel deployment is performed | ||
306 | without Internet connectivity. The <emphasis | ||
307 | role="bold">tools</emphasis> Puppet task has a 5 minute completion | ||
308 | timeout, causing there to not be enough time left for the other | ||
309 | components of the tools task to be executed.</para> | ||
310 | </listitem> | ||
311 | |||
312 | <listitem> | ||
313 | <para>Workaround:</para> | ||
314 | |||
315 | <para>Simply press <literal>Deploy changes</literal> again. Since | ||
316 | by now Postfix is installed, it will skip the installation and | ||
317 | finish in time.</para> | ||
318 | </listitem> | ||
319 | </itemizedlist> | ||
320 | </listitem> | ||
321 | </orderedlist> | ||
322 | </section> | ||
323 | |||
324 | <section condition="hidden" id="bugs-limitations-doc"> | ||
325 | <title>Documentation</title> | ||
326 | |||
327 | <itemizedlist spacing="compact"> | ||
328 | <listitem> | ||
329 | <para><emphasis role="bold">PDF navigation</emphasis>: When using | ||
330 | links to open other PDFs, or jump to another place in the same PDF, | ||
331 | jumping back sometimes fails. This has been observed when opening a | ||
332 | PDF in Adobe Reader, inside a browser with PDF add-on, as well as when | ||
333 | the browser is configured to open PDF files in an external PDF reader. | ||
334 | As a workaround, open the HTML version of the | ||
335 | document.<remark>LXCR-3283</remark></para> | ||
336 | </listitem> | ||
337 | </itemizedlist> | ||
338 | </section> | ||
339 | |||
340 | <section condition="hidden" id="bugs-limitations-other"> | ||
341 | <title>Miscellaneous</title> | ||
342 | |||
343 | <itemizedlist spacing="compact"> | ||
344 | <listitem> | ||
345 | <para></para> | ||
346 | |||
347 | <remark>list all misc problems here.</remark> | ||
348 | </listitem> | ||
349 | </itemizedlist> | ||
350 | </section> | ||
351 | |||
352 | <!-- The file with a section below is autocreated by make init | ||
353 | |||
354 | <xi:include href="jiraissues_generated.xml" | ||
355 | xmlns:xi="http://www.w3.org/2001/XInclude" /> --> | ||
356 | </chapter> \ No newline at end of file | ||