2 Chapter 32. Multitape support in Amanda 2.2
3 Prev Part VI. Historical files Next
5 -------------------------------------------------------------------------------
7 Chapter 32. Multitape support in Amanda 2.2
18 XML-conversion;Updates
54 Refer to http://www.amanda.org/docs/multitape.html for the current version of
59 The goal of this enhancement is to make Amanda independent of the number of
60 tapes used per run or even per dump cycle. Specifically, I would like Amanda to
63 * output of amdump run goes to more than one tape
64 * a single dump file can straddle two tapes
65 * more than one amdump run can be done in a single day
66 * planner should not care how many runs per cycle occur
70 * multiple runs of amdump can go onto one tape (eg an append mode)
71 * any dump files from a previous run that are on the holding disk are written
72 to tape in this run (eg eliminate amflush)
73 * taper write to multiple tape drives simultaneously
81 Previously, planner marked time by the number of amdump runs, which it equated
82 with number of tapes, and number of days. In Amanda 2.2, Amanda keeps track of
83 the real passage of time, and doesn't generally care about the number of runs
84 or tapes between any two events.
85 While Amanda 2.2 doesn't care about spacing between runs, dump cycles are still
86 in terms of days, to make things easy to understand for the user. So, time
87 differences are rounded to the nearest 24 hours:
88 days_diff(A,B) = (<B> - <A> + 86400/2) / 86400
89 Where the times A and B are in seconds since the Unix epoch, and 86400 is the
90 number of seconds per day. This rounds a 2.49 day difference down to 2 days,
91 and a 2.5 day difference up to 3 days. No, Olafur, Unix time does not handle
92 leap seconds. Give me a break. :-)
96 The first thing planner does is calculate when each filesystem is due for a
97 full backup. This is trivial for normal backups:
98 full_is_due = days_diff(<time of last full>, <curtime>) >= dumpcycle
99 There is a complication for "skip-full" filesystems. Under 2.2, these will be
100 skipped on any runs that occur on the day the full is due, but we have to do
101 the right thing if multiple runs are done that day, and if no runs are done
102 that day (in which case we should be doing an incremental). Also, the time of
103 last full dump is a fiction maintained by the planner -- Amanda has no way to
104 tell whether the full backup was actually done or when it was done:
108 <time of last full> += dumpcycle;
109 if(days_diff(<time of last full>, <curtime>) == 0)
110 skip the filesystem on this run;
112 do an incremental dump of this filesystem;
119 The runtapes parameter tells planner how many tapes it should plan to use each
120 run. It multiplies this by the tape length to get the size available for the
121 run. (NOTE: later amend this size if appending to tapes, or if there are dumps
122 on the holding disk waiting to be flushed). Other than the size calculation,
123 planner doesn't really care how many tapes will be written to.
124 The fundamental problem with attempting to balance the schedule is that we no
125 longer know how many amdump runs will be done in a full cycle. The number may
126 change from cycle to cycle if there are extenuating circumstances.
127 So, planner must guess at how many runs will be done in one cycle, by looking
128 at the information for the last cycle, or, if this is the first cycle, assuming
129 one run for each day in the dump cycle.
133 When can a tape be overwritten, considering that it might have old dumps on it?
134 We want to be able to warn when full dumps are going to be overwritten, but
135 given the possibility of old files on the tape, how can we know when the tape
136 is no longer needed? I think we can get this when going through the info file,
137 considering each full dump and what tape it is on. Make sure we correctly
138 handle stale information.
145 taper must now handle writing to multiple tapes in one night, but choosing the
146 tapes from the tape rack is done one at a time as needed, re-applying the same
147 algorithm each time (see Amanda_Tape_Changer_Support).
151 As in earlier versions of Amanda, taper itself does not try to restrict writing
152 to the tape size given in the config file. It relied on planner having
153 correctly estimated backup sizes and limiting itself to what would fit on one
155 Now, taper needs to switch to a new tape when the current tape has filled up.
156 The tape is considered full when taper gets a write error. This will most
157 likely occur in the middle of writing a (potentially large) backup file,
158 perhaps even from a direct-to-tape socket, so there is no possibility of
159 starting the backup file over again on the next tape, it must start from where
160 it left off, rewriting the block that got the error on the next tape.
161 To insure correct operation, the file header of the continued file should
162 contain an indication that it is a continuation, and at what offset. amrestore
163 of course needs to be aware of this scheme and handle it correctly, perhaps by
164 double-buffering internally. XXX provide more alg details here, or just leave
165 it with the general idea?
169 We need to specify the sequence number of the tape in the run, in the tape
170 header file. The file header block specifies whether it is a continuation file
173 Tapelist File Changes
175 The lines in the tapelist file should contain the sequence number of the tape
176 in its run, as well as the amount of data written on the tape, and perhaps
177 whether or not the end of tape was reached.
178 -------------------------------------------------------------------------------
181 Chapter 31. What once was new Home Chapter 33. Thoughts about a Strategy API