source: gpfs_3.1_ker2.6.20/share/man/man8/mmfileid.8 @ 173

Last change on this file since 173 was 16, checked in by rock, 17 years ago
File size: 6.2 KB
Line 
1.TH  mmfileid 03/02/06
2mmfileid Command
3.SH "Name"
4.PP
5\fBmmfileid\fR - Determines which files are located on areas of a disk
6that are damaged or considered to be suspect.
7.PP
8\fBAttention:\fR Use
9this command only when directed by the IBM Support Center.
10.PP
11\fB
12Synopsis\fR
13.PP
14\fBmmfileid\fR \fIDevice\fR {\fB
15-d\fR [\fINodeName\fR]:{\fIDiskName\fR|\fIDiskNum\fR|\fB
16BROKEN\fR}:[\fIPhysAddr1\fR[-\fIPhysAddr2\fR]] | \fB
17-F\fR \fIDescFile\fR} [\fB-o\fR \fIOutputFile\fR]
18[\fB-f\fR \fINumThreads\fR] [\fB-t\fR \fIDirectory
19\fR]
20.PP
21\fBDescription\fR
22.PP
23Use the \fBmmfileid\fR command to determine which files are located on areas of a disk
24that are damaged or considered to be suspect.
25Before running \fBmmfileid\fR, the user must run a disk analysis utility
26and obtain the disk sector numbers that are suspect or known to be damaged.
27These sectors are input to the \fBmmfileid\fR command.
28To run the \fBmmfileid\fR command, you must specify the physical volume name of the disk,
29the GPFS disk ID number, or the keyword \fBBROKEN\fR to find all
30files with broken addresses resulting in the loss of data.
31The output of the command is a list of inodes and names of files that are located on the input
32disk sectors.
33.PP
34The output can be redirected to
35a file (using the \fB-o\fR flag) and sorted on the inode number, using the \fBsort\fR command.
36The \fBmmfileid\fR command output contains one line for each inode found to be located
37on the corrupt disk sector. The file name is relative to the root of the file system in which
38it resides. The command output has this format:
39.sp
40.nf
41inode_number   Logical_Disk_Address    snapshotid    filename
42.fi
43.sp
44.PP
45Only one instance of \fBmmfileid\fR may be active on
46a GPFS cluster at any given time. The \fBmmfileid\fR command will
47not work for a file system owned and controlled by another GPFS cluster.
48.PP
49\fB
50Parameters\fR
51.PP
52.RS +3
53\fB\fIDevice\fR\fR
54.RE
55.RS +9
56The device name for the file system on which this utility is to be run. 
57This must be the first parameter and is required.
58.RE
59.PP
60.RS +3
61\fB\fB-d\fR [\fINodeName\fR]:{\fIDiskName\fR|\fI
62DiskNum\fR|\fBBROKEN\fR}:[\fIPhysAddr1\fR[-\fIPhysAddr2
63\fR]]
64.PP
65\fR
66.RE
67.RS +9
68.PP
69.RS +3
70\fB\fINodeName\fR\fR
71.RE
72.RS +9
73Specifies a node in the GPFS cluster that has access to the disk in question.
74.RE
75.PP
76.RS +3
77\fB\fIDiskName\fR\fR
78.RE
79.RS +9
80Specifies the physical volume name of the disk.
81.RE
82.PP
83.RS +3
84\fB\fIDiskNum\fR\fR
85.RE
86.RS +9
87Specifies the GPFS disk ID number of the disk as currently displayed by
88the \fBmmlsdisk -L\fR command.
89.RE
90.PP
91.RS +3
92\fBBROKEN\fR
93.RE
94.RS +9
95Specifies to find all files that have broken addresses resulting in lost
96data.
97.RE
98.PP
99.RS +3
100\fB\fIPhysAddr1\fR[-\fIPhysAddr2\fR]\fR
101.RE
102.RS +9
103Specified the range of physical disk addresses, as reported by the error
104logging utility, for the disk that has encountered a failure.
105The default value for \fIPhysAddr1\fR is zero.
106The default value for \fIPhysAddr2\fR is the value for \fIPhysAddr1\fR.
107.PP
108If both \fIPhysAddr1\fR and \fI
109PhysAddr2\fR are zero, the entire disk is searched.
110.RE
111.RE
112.PP
113.RS +3
114\fB-F \fIDescFile\fR\fR
115.RE
116.RS +9
117Specifies a file containing a list of disks and corresponding disk addresses,
118for the file system on which the \fBmmfileid\fR command is to be
119run. The file uses the same conventions as the \fB-d\fR option.
120.PP
121The format of the file is:
122.sp
123.nf
124#optional comment line
125nodename1:hdisk1:PhysAddr1-PhysAddr2
126nodename1:hdisk1:PhysAddr3-PhysAddr4
127nodename3:hdisk2:PhysAddr1-PhysAddr2
128nodename1:hdiskn:PhysAddr1-PhysAddr2
129.fi
130.sp
131.PP
132For example:
133.sp
134.nf
135:gpfs1nsd
136:10:27645856
137:BROKEN
138.fi
139.sp
140.RE
141.PP
142.RS +3
143\fB-f \fINumThreads\fR\fR
144.RE
145.RS +9
146Specifies the number of worker threads that are to be created by the \fB
147mmfileid\fR command.
148.PP
149The default is value 16. The minimum value is
1501. If \fInumThreads\fR is less than 2, only the main process performs
151work. The maximum value can be as large as is allowed by the operating system \fB
152pthread_create\fR function for a single process. A suggested value is
153twice the number of disks in the file system. If an error occurs creating
154a worker thread, \fBmmfileid\fR exits with an error.
155.RE
156.PP
157.RS +3
158\fB-o \fIOutputFile\fR\fR
159.RE
160.RS +9
161Specifies the name of a file for the \fBmmfileid\fR command
162to write its output. If not specified, output is sent to standard output.
163.RE
164.PP
165.RS +3
166\fB-t \fIDirectory\fR\fR
167.RE
168.RS +9
169Specifies the directory to use for the \fBmmfileid\fR temporary
170files and the sort program intermediate files. The default directory is \fB
171/tmp\fR.
172.RE
173.PP
174\fBExit status\fR
175.PP
176.PP
177.RS +3
178\fB0\fR
179.RE
180.RS +9
181Successful completion.
182.RE
183.PP
184.RS +3
185\fBnonzero\fR
186.RE
187.RS +9
188A failure has occurred.
189.RE
190.PP
191\fBSecurity\fR
192.PP
193You must have root authority to run
194the \fBmmfileid\fR command.
195.PP
196This command must be invoked
197on a file system that is not running an excessive number of reads and writes.
198Excessive file system activity will produce incorrect results.
199.PP
200\fB
201Examples\fR
202.PP
203Assume that a disk analysis tool reported that \fB
204hdisk6\fR, \fBhdisk7\fR, \fBhdisk8\fR, and \fB
205hdisk9\fR contained bad sectors. This command:
206.sp
207.nf
208mmfileid /dev/gpfsB -F addr.in
209.fi
210.sp
211.PP
212where \fBaddr.in\fR contains this:
213.sp
214.nf
215k148n07:hdisk9:2206310-2206810
216k148n07:hdisk8:2211038-2211042
217k148n07:hdisk8:2201800-2202800
218k148n01:hdisk6:2921879-2926880
219k148n09:hdisk7:1076208-1076610
220.fi
221.sp
222.PP
223produces output similar to this:
224.sp
225.nf
226Address 2201958 is contained in the Block allocation map (inode 1)
227Address 2206688 is contained in the ACL Data file (inode 4, snapId 0)
228Address 2211038 is contained in the Log File (inode 7, snapId 0)
22914336 1076256 0 /gpfsB/tesDir/testFile.out
23014344 2922528 1 /gpfsB/x.img
231.fi
232.sp
233.PP
234The lines starting with Address are special GPFS files. If your
235output contains any of these lines, \fBdo not\fR attempt to replace or
236repair the indicated files.
237Call the IBM Support Center for assistance.
238This output indicates that files \fB/gpfsB/tesDir/testFile.out\fR and \fB
239/gpfsB/x.img\fR are located on bad sector areas.
240.PP
241The format for
242each line of output is as follows:
243.sp
244.nf
245inode_number   Logical_Disk_Address    snapshotid    filename
246.fi
247.sp
248.PP
249\fBSee also\fR
250.PP
251The section on \fIDisk media
252failure\fR in \fIGeneral Parallel File System: Problem
253Determination Guide\fR.
254.PP
255\fBLocation\fR
256.PP
257\fB
Note: See TracBrowser for help on using the repository browser.