"/var/tmp/ExNbaObk" 设备上没有剩余空间 如何解决?急

leekingli

我认为是你的系统盘除了问题.

leekingli

我认为是你的系统盘出了问题.

heeh

\"/var/tmp/Exp2aOeI\" 设备上没有剩余空间\r\n\r\n# chmod 1777 /var/tmp\r\n# ls -la /var/tmp\r\n总数 6\r\ndrwxrwxrwt   2 root     sys         2048  3月  6 18:41 .\r\ndrwxr-xr-x  35 root     sys          512 2004   5月  8 ..\r\n# ls -l /var/tmp\r\n总数 0\r\n# ls -ald /var /var/tmp\r\ndrwxr-xr-x  35 root     sys          512 2004   5月  8 /var\r\ndrwxrwxrwt   2 root     sys         2048  3月  6 18:41 /var/tmp\r\n# vi /etc/hosts\r\n\"/var/tmp/Exk6aiiI\" 设备上没有剩余空间\r\n\r\n# mount |grep /var\r\n/var on /dev/dsk/c1t0d0s4 read/write/setuid/intr/largefiles/logging/xattr/onerror=panic/dev=1d80004 on 二  3月  6 18:43:02 2007\r\n/var/run on swap read/write/setuid/xattr/dev=1 on 二  3月  6 18:43:02 2007\r\n# swap -l\r\n交换文件             dev  swaplo blocks   free\r\n/dev/dsk/c1t0d0s1   118,1      16 16617392 16617392\r\n# df -k\r\n文件系统              千字节    用了    可用 容量      挂接在\r\n/dev/dsk/c1t0d0s0    16967594 5086177 11711742    31%    /\r\n/dev/dsk/c1t0d0s3    5046030 1616767 3378803    33%    /usr\r\n/proc                      0       0       0     0%    /proc\r\nmnttab                     0       0       0     0%    /etc/mnttab\r\nfd                         0       0       0     0%    /dev/fd\r\n/dev/dsk/c1t0d0s4    10086988 1555475 8430644    16%    /var\r\nswap                 11695880      24 11695856     1%    /var/run\r\nswap                 11695856       0 11695856     0%    /tmp\r\n/dev/dsk/c1t0d0s6    20169673 3567177 16400800    18%    /bind\r\n/dev/dsk/c1t0d0s5    10086988   48954 9937165     1%    /opt\r\n# cat /etc/vfstab\r\n#device         device          mount           FS      fsck    mount   mount\r\n#to mount       to fsck         point           type    pass    at boot options\r\n#\r\nfd      -       /dev/fd fd      -       no      -\r\n/proc   -       /proc   proc    -       no      -\r\n/dev/dsk/c1t0d0s1       -       -       swap    -       no      -\r\n/dev/dsk/c1t0d0s0       /dev/rdsk/c1t0d0s0      /       ufs     1       no      logging\r\n/dev/dsk/c1t0d0s3       /dev/rdsk/c1t0d0s3      /usr    ufs     1       no      logging\r\n/dev/dsk/c1t0d0s4       /dev/rdsk/c1t0d0s4      /var    ufs     1       no      logging\r\n/dev/dsk/c1t0d0s6       /dev/rdsk/c1t0d0s6      /bind   ufs     2       yes     logging\r\n/dev/dsk/c1t0d0s5       /dev/rdsk/c1t0d0s5      /opt    ufs     2       yes     logging\r\n#/dev/md/dsk/d3 /dev/md/rdsk/d3 /usr    ufs     1       no      logging \r\n#/dev/md/dsk/d4 /dev/md/rdsk/d4 /var    ufs     1       no      logging \r\n#/dev/md/dsk/d6 /dev/md/rdsk/d6 /bind   ufs     2       yes     logging \r\n#/dev/md/dsk/d5 /dev/md/rdsk/d5 /opt    ufs     2       yes     logging \r\nswap    -       /tmp    tmpfs   -       yes     -\r\n\r\n\r\n\r\n我发现vfstab中的logging有些多..跟这还有关系吗?客户最近才发现的这问题.

heeh

查看过messages里面有一个报错..\r\nMar  6 18:45:00 sun480R ufs: [ID 682040 kern.notice] NOTICE: /var: out of inodes\r\n听SUN工程师说,只能备份 /var ,then newfs /var   restore /var

chinaux

you might remove those small files if not needed, or, migrate the directories of small files to another file systems, and create a symblic link.

heeh

原帖由 chinaux 于 2007-4-18 13:42 发表\r\nyou might remove those small files if not needed, or, migrate the directories of small files to another file systems, and create a symblic link.

\r\n你的意思是这样?\r\n#rm -rf /var/tmp\r\n#mkdir /opt/tmp\r\n#chmod 0755 /opt/tmp \r\n#chown root:sys /opt/tmp \r\n#ln -s /var/tmp /opt/tmp\r\n\r\n我作了.有问题

chinaux

原帖由 heeh 于 2007-4-18 13:46 发表\r\n\r\n你的意思是这样?\r\n#rm -rf /var/tmp\r\n#mkdir /opt/tmp\r\n#chmod 0755 /opt/tmp \r\n#chown root:sys /opt/tmp \r\n#ln -s /var/tmp /opt/tmp\r\n\r\n我作了.有问题

\r\n\r\nNo.\r\n\r\nThe issue is that \"inodes are used up on /var\". So, there must be a great deal of small files on the file system, probably, they are under /var/spool/mqueue. \r\n\r\nAssuming /var/spool/mqueue has a lot of small files. In order to migrate /var/spool/mqueue, you might do the following:\r\n1. bring down sendmail\r\n2. mkdir -p /opt/var.spool.mqueue\r\n3. cd /var/spool/mqueue; tar cf - . | (cd /opt/var.spool.mqueue; tar xf -)\r\n4. verify the data migration.\r\n5. if no issues,  rm -r /var/spool/mqueue; ln -s /opt/var.spool.mqueue /var/spool/mqueue\r\n6. start sendmail

chinaux

you might have to find out how much inodes are needed for migrating a directory and how much inodes are available on the potential destinations:\r\n\r\ndf -F ufs -o i

susbin

You might want to boot the system from a Solaris CD and run \r\nfsck \r\non /var.\r\n\r\nIt looks like that many files are deleted under /var but the inodes of them were not released yet.\r\n\r\nTo free up some disk space under /var, \r\nyou can check these directories see if any files can be deleted or moved to somewhere else. \r\n\r\n/var/cron/log\r\n\r\n/var/spool/lp/logs\r\n\r\n/var/adm/utmp\r\n\r\n/var/adm/utmpx\r\n\r\n/var/adm/wtmp\r\n\r\n/var/adm/wtmpx\r\n\r\n/var/log/syslog*\r\n\r\n/var/adm/messages.*