1
2
3 :Info: raw_access.gi: 02/26/82 raw access
4
5 Also referred to as the raw mode, it is the access mode granted a
6 process to an object by discretionary access control. Raw access to
7 an object is computed from the access control list ACL, ring
8 brackets, and AIM attributes of the object. See also access.info.
9
10
11 :Info: ready_message.gi: 02/26/82 ready message
12
13 A message that is printed each time a user is at command level.
14 Printing this message may be inhibited, or the user may define his or
15 her own ready message. The standard system ready message tells the
16 time of day, the number of CPU seconds and pages of information
17 brought into main memory since the last ready message, and the current
18 listener level if greater than 1. See the general_ready, ready_on
19 and ready_off commands.
20
21
22 :Info: record.gi: 02/26/82 record
23
24 The term "record" has two meanings on Multics. In one sense, it
25 means the smallest unit of disk allocation, containing 1024 36-bit
26 words 4096 characters. In PL/I and FORTRAN, it means a block of
27 data transferred during input or output.
28
29
30 :Info: recursion.gi: 02/26/82 recursion
31
32 The ability of a procedure to invoke itself.
33
34
35 :Info: reference_name.gi: 02/26/82 reference name
36
37 When a segment is made known to a process, particular names may
38 be associated with it in that process. This is called initiation.
39 Thereafter, a symbolic reference to this reference name is directed to
40 the associated segment. Reference names need not be the same as any
41 of the segment's entrynames.
42
43
44 :Info: relative_pathname.gi: 02/26/82 relative pathname
45
46 A pathname that names a segment in its relation to the working
47 directory.
48
49
50 :Info: retrieval.gi: 02/10/84 retrieval
51
52 The process of copying a segment or directory back into the storage
53 system hierarchy from backup tapes. Multics has two backup systems:
54 volume and hierarchy. These backup systems copy segments and
55 directories from the storage system hierarchy onto tapes at specified
56 intervals. When you ask for a retrieval, you are asking the operator
57 to copy your segment or directory from these tapes back into the
58 storage system hierarchy.
59
60 The operator can use tapes produced by either backup system to
61 recover segments and directories. But the primary functions of the
62 two kinds of backup tapes are different. See "Notes on backup tapes"
63 next.
64
65
66 Notes on backup tapes:
67
68 Volume backup tapes are primarily used for recovery from major
69 system failures which damage a large portion of the file system. An
70 example of this is a head crash which causes physical damage to a disk
71 pack. When this happens, all of the segments which reside on the
72 damaged disk pack must be restored onto a new pack. Segments are
73 written to volume backup tapes at frequent intervals. However, these
74 tapes are normally retained only for short periods of time, perhaps
75 two or three weeks, depending on site policy.
76
77 Hierarchy backup tapes are primarily used for long term backup
78 storage. Segments are written to hierarchy backup tapes at less
79 frequent intervals. However, these tapes are normally retained for
80 long periods of time, perhaps a year or more, again depending on site
81 policy.
82
83
84 The operator normally uses hierarchy backup tapes to recover
85 individual segments and directories that you yourself have deleted or
86 which have been damaged during a system crash hardware software or
87 power failure.
88
89 In addition to the fact that hierarchy tapes are retained longer,
90 there is another major reason why they are usually used for individual
91 retrievals. User segments are in sequential order on hierarchy tapes.
92 They are not in sequential order on volume tapes, and in fact, may be
93 spread over MANY tapes. Thus, fewer hierarchy tapes than volume tapes
94 need to be mounted to do a retrieval.
95
96
97 The procedure for requesting a retrieval varies. Your site should
98 have its own help file spelling out its policy on requesting
99 retrievals and explaining its schedule of when backup tapes are
100 created. This help file should be kept in the directory called
101 >doc>installation_info_segs >doc>iis. If your site doesn't have a
102 help file like this, please check with your System Administrator for
103 the correct, exact procedure to follow at your site. Usually, the
104 procedure involves sending mail to the operator. See "Notes on
105 sending mail to operator" next.
106
107
108 Notes on sending mail to operator:
109
110 When you send mail to the operator, you should include the complete
111 pathnames of the segments/directoryies you want retrieved, and
112 the approximate date and time they were last modified.
113
114 Be sure to type the pathnames correctly, using capital letters when
115 required. Use primary names of segments and directories, not
116 add_names, because that makes it easier for the operator to find your
117 files.
118
119 If you are not certain of the date and time the version of the
120 segment/directory you want to retrieve was modified, give the operator
121 a range of dates and times within which you believe it was modified
122 e.g.: modified between May 16 and June 1; modified yesterday between
123 1230 and 1400. See "Examples of sending mail to operator" below.
124
125
126 The operator will determine which if any backup system produced
127 the tape copy of the segment or directory you want to retrieve.
128 Whether or not your segment/directory can be retrieved depends on when
129 and how long it was on the system, and what the backup schedule in
130 place at your site is like.
131
132 To understand the backup schedule, you need to know more about the
133 backup systems. See "Notes on backup systems" next.
134
135
136 Notes on backup systems:
137
138 The procedure by which the volume and hierarchy backup systems copy
139 segments and directories from the storage system hierarchy onto tape
140 is called dumping. The segments and directories selected for dumping
141 are determined by the mode -- incremental, consolidated, or complete
142 -- in which the dumping is performed.
143
144
145 An incremental dump locates and copies all segments and directories
146 which have been created or modified since the last time an incremental
147 dump was done. Its main function is to limit the amount of
148 information that can be lost due to changes that have been made since
149 the last incremental tape was created.
150
151 A consolidated dump locates and copies all segments and directories
152 which have been modified after some specified time in the past. Its
153 main function is to consolidate the most recent information stored on
154 a group of incremental tapes, thus decreasing the number of tapes
155 which must be saved and processed. It does this by copying all
156 segments which have been modified since the last consolidated dump.
157
158
159 A complete dump locates and copies every segment and directory in
160 the storage system without regard for when they were last modified.
161 Its main function is to establish a checkpoint in time. If it's ever
162 necessary to recover a major portion of online storage, the tape with
163 the most recent complete dump on it marks a cutoff point, beyond which
164 no older dump tapes are needed.
165
166 Now you have enough background to understand a backup schedule.
167 See "Examples of backup schedule" next.
168
169
170 Examples of backup schedule:
171
172 The following example shows a schedule of when different backup
173 procedures could be run and how long the tapes they produce could be
174 held. Keep in mind that some sites only run one of the backup
175 systems, not both of them. Also keep in mind that segments and
176 directories are only copied to incremental and consolidated backup
177 tapes when they are newly created or when they have been modified.
178
179
180 procedure interval retention
181
182 Hierarchy Incremental every 12 hours 2 weeks
183
184 Hierarchy Consolidated every Saturday 1 year
185
186 Hierarchy Complete end of month 1 year
187 normally between
188 26th and 30th
189
190 Volume Incremental hourly 7 days
191
192 Volume Consolidated daily 1800 hrs 7 days
193
194 Volume Complete 7 days 14 days
195
196
197 Notes on backup schedule:
198
199 The first three procedures shown in the example provide long term
200 backup protection of the file system. The tapes they produce are the
201 ones the operator normally uses for retrieving user segments and
202 directories.
203
204 The last three procedures shown in the example provide a method of
205 recovering from major system crashes which damage the file system.
206 The tapes they produce are mainly used for rebuilding physical volumes
207 disk packs after they are destroyed by head crashes and have to be
208 replaced. However, the tapes they produce can also be used to recover
209 individual segments or directories.
210
211
212 Examples of sending mail to operator:
213
214 sdm Operator
215 Subject: retrieval request
216 Message:
217 Please retrieve the following files:
218
219
220 1. >udd>Mcc>Johnson>test_A.fortran
221 modified yesterday morning
222 between 800 and 1000
223
224 2. >udd>Mcc>Johnson>file_B
225 modified 1st week in February and
226 deleted from system sometime in Apr.
227
228 3. >udd>Mcc>Johnson>file_C
229 created four weeks ago last Tuesday and
230 deleted the following day.
231
232 4. >udd>Mcc>Johnson>PLI directory
233 as it was the end of Nov - 9 months ago.
234 .
235
236
237 If the schedule shown above under "Examples of backup schedule" was
238 the one being used, the operator would be able to retrieve all of the
239 segments and directories in this example except for # 3. Since # 3
240 was only on the system for one day, it was only captured on short term
241 retention tapes that are overwritten in 2 to 3 weeks. If the
242 retrieval request had been made within two weeks, the operator would
243 have been able to retrieve the segment.
244
245
246 If your site allows users to use the enter_retrieval_request
247 command, you should refer to the description of that command.
248 Remember that you can only use it to do a VOLUME retrieval. It
249 doesn't work for a hierarchy retrieval.
250
251
252 :Info: ring.gi: 02/26/82 ring
253
254 A particular level of privilege at which programs may execute.
255 Lower numbered rings are of higher privilege than higher numbered
256 ones. The supervisor program runs in ring 0, most user programs run
257 in ring 4.
258
259
260 :Info: ring_brackets.gi: 02/26/82 ring brackets
261
262 A set of integers associated with each segment that define in
263 what rings that segment may be written, read, called, or executed.
264
265
266 :Info: ring_structure.gi: 03/05/82 ring structure
267
268 The structure of access control on Multics which is implemented
269 by special hardware. Operation is controlled in such a way that the
270 computer's work is done in a number of mutually exclusive subsets.
271 These subsets may be considered concentric rings of privilege,
272 representing different levels of access rights. The innermost or
273 hardcore ring is made up of those segments essential to all users.
274 This innermost ring, designated as ring 0, represents the highest
275 level of privilege. The work of most users is done in ring 4. Ring 7
276 is the ring of least privilege.
277
278
279 :Info: root.gi: 02/26/82 root
280
281 The directory that is the base of the directory hierarchy. All
282 other directories are subordinate to it. It has an absolute pathname
283 of >.