Details
-
Bug
-
Status: Closed (View Workflow)
-
Major
-
Resolution: Fixed
-
10.2.7
-
None
Description
On a potentially corrupted server instance mariabackup fails with errno 14 EFAULT - Bad Address
mariabackup --backup --user=root --password=... --target-dir=.../backup/$(date +\%Y-\%m-\%d_\%H:\%M:\%S)
|
170728 15:56:46 Connecting to MySQL server host: localhost, user: root, password: set, port: 3306, socket: /var/run/mysqld/mysqld.sock
|
Using server version 10.2.7-MariaDB-10.2.7+maria~xenial-log
|
mariabackup based on MariaDB server 10.2.7-MariaDB Linux (x86_64)
|
xtrabackup: uses posix_fadvise().
|
xtrabackup: cd to /var/lib/mysql/
|
xtrabackup: open files limit requested 215200, set to 215200
|
xtrabackup: using the following InnoDB configuration:
|
xtrabackup: innodb_data_home_dir = .
|
xtrabackup: innodb_data_file_path = ibdata1:12M:autoextend
|
xtrabackup: innodb_log_group_home_dir = ./
|
2017-07-28 15:56:46 140398601803584 [Note] InnoDB: Number of pools: 1
|
mariabackup: Error writing file '.../backup/2017-07-28_15:56:46/ib_logfile0' (Errcode: 14 "Bad address")
|
xtrabackup: Error: write to logfile failed
|
xtrabackup: Error: xtrabackup_copy_logfile() failed.
|
strace shows that after a few small reads from input file ib_logfile0 input file iblogfile1 is opened, 64KB are read from it, and an attempt to write ~18 ExaByte
to the ib_logfile0 output file in the backup target dir is made.
mkdir("/mysqldepot/backup/2017-07-29_16:47:46/", 0700) = 0
|
open("./ib_logfile0", O_RDONLY) = 4
|
pread64(4, "\0\0\0\1\0\0\0\0\0\0\ft\370\36\200\0MariaDB 10.2.7\0\0"..., 512, 0) = 512
|
pread64(4, "\0\0\0\0\0\4\373^\0\0\fx\237\1C\364\0\0\0\3\246\342\323\364\0\0\0\0\4\0\0\0"..., 512, 512) = 512
|
pread64(4, "\0\0\0\0\0\4\373_\0\0\fx\237\1D\r\0\0\0\3\246\342\324\r\0\0\0\0\4\0\0\0"..., 512, 1536) = 512
|
pread64(4, "\0\0\0\1\0\0\0\0\0\0\ft\370\36\200\0MariaDB 10.2.7\0\0"..., 512, 0) = 512
|
pread64(4, "\0\0\0\0\0\4\373^\0\0\fx\237\1C\364\0\0\0\3\246\342\323\364\0\0\0\0\4\0\0\0"..., 512, 512) = 512
|
pread64(4, "\0\0\0\0\0\4\373_\0\0\fx\237\1D\r\0\0\0\3\246\342\324\r\0\0\0\0\4\0\0\0"..., 512, 1536) = 512
|
mkdir("/mysqldepot/backup/2017-07-29_16:47:46/", 0700) = -1 EEXIST (File exists)
|
mkdir("/mysqldepot/backup/2017-07-29_16:47:46/", 0700) = -1 EEXIST (File exists)
|
open("/mysqldepot/backup/2017-07-29_16:47:46/ib_logfile0", O_WRONLY|O_CREAT|O_EXCL|O_NOFOLLOW, 0660) = 5
|
write(5, "\0\0\0\1\0\0\0\0\0\0\fx\237\1D\rBackup 10.2.7-Ma"..., 512) = 512
|
fadvise64(5, 0, 0, POSIX_FADV_DONTNEED) = 0
|
write(5, "\0\0\0\0\0\4\373_\0\0\fx\237\1D\r\0\0\0\0\0\0\10\r\0\0\0\0\4\0\0\0"..., 512) = 512
|
fadvise64(5, 0, 0, POSIX_FADV_DONTNEED) = 0
|
write(5, "\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"..., 512) = 512
|
fadvise64(5, 0, 0, POSIX_FADV_DONTNEED) = 0
|
write(5, "\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"..., 512) = 512
|
fadvise64(5, 0, 0, POSIX_FADV_DONTNEED) = 0
|
open("./ib_logfile1", O_RDONLY) = 6
|
pread64(6, "\274O\200\243\0\26\0\r\0\4\373_\3648\0\0\fx\237\1D\r\0\0\0\0\0\0\0\0\0\0"..., 65536, 7094850560) = 65536
|
write(5, "\274O\200\243\0\26\0\r\0\4\373_\3648\0\0\fx\237\1D\r\0\0\0\0\0\0\0\0\0\0"..., 18446730364173942784) = -1 EFAULT (Bad address)
|
write(2, "\7", 1) = 1
|
write(2, "mariabackup", 11) = 11
|
write(2, ": ", 2) = 2
|
write(2, "Error writing file '/mysqldepot/"..., 99) = 99
|
write(2, "\n", 1) = 1
|
write(2, "xtrabackup: Error: write to logf"..., 43) = 43
|
From looking at the error message text this seems to be happing in xtrabackup_scan_log_recs() in extra/mariabackup/xtrabackup.cc
So either we have something overwriting the write_size variable there, giving it an insanely large value, or the values read and used to calculate the write_size are already corrupted to begin with.
If the later is true the some sanity checks should be added to catch values that exceed innodb_log_file_size and report some more meaningful error message in that case ...
Attachments
Issue Links
- is duplicated by
-
-
- Closed
-
-
-
- Closed
-
- relates to
-
-
- Closed
-