I'm slowly getting everything back to normal here after swapping motherboards on my server. I also took the opportunity to bring the server up to CentOS 5. I'm still getting services working but I have one oddity with amanda. Ever since I upgraded the server to CentOS 5, amcheck dies and gives me a backtrace if there are no available tapes. amcheck works fine if the expected next tape is in my tape changer. If it's time to change tapes, I get:
*** glibc detected *** /usr/sbin/amcheck: double free or corruption (fasttop): 0x09fb1a28 *** ======= Backtrace: ========= /lib/libc.so.6[0x9e1aa6] /lib/libc.so.6(cfree+0x90)[0x9e4fc0] /usr/sbin/amcheck[0x820e03] /usr/sbin/amcheck(main+0xc2d)[0x82291d] /lib/libc.so.6(__libc_start_main+0xdc)[0x990dec] /usr/sbin/amcheck[0x81da01] ======= Memory map: ======== 00110000-00135000 r-xp 00000000 09:02 2378087 /lib/libm-2.5.so 00135000-00136000 r-xp 00024000 09:02 2378087 /lib/libm-2.5.so 00136000-00137000 rwxp 00025000 09:02 2378087 /lib/libm-2.5.so 00137000-00140000 r-xp 00000000 09:02 2378088 /lib/libnss_files-2.5.so 00140000-00141000 r-xp 00008000 09:02 2378088 /lib/libnss_files-2.5.so 00141000-00142000 rwxp 00009000 09:02 2378088 /lib/libnss_files-2.5.so 002cd000-002d8000 r-xp 00000000 09:02 2378249 /lib/libgcc_s-4.1.2-20070626.so.1 002d8000-002d9000 rwxp 0000a000 09:02 2378249 /lib/libgcc_s-4.1.2-20070626.so.1 0046b000-0047a000 r-xp 00000000 09:02 2378128 /lib/libresolv-2.5.so 0047a000-0047b000 r-xp 0000e000 09:02 2378128 /lib/libresolv-2.5.so 0047b000-0047c000 rwxp 0000f000 09:02 2378128 /lib/libresolv-2.5.so 0047c000-0047e000 rwxp 0047c000 00:00 0 006ab000-006be000 r-xp 00000000 09:02 2378125 /lib/libnsl-2.5.so 006be000-006bf000 r-xp 00012000 09:02 2378125 /lib/libnsl-2.5.so 006bf000-006c0000 rwxp 00013000 09:02 2378125 /lib/libnsl-2.5.so 006c0000-006c2000 rwxp 006c0000 00:00 0 0081c000-00825000 r-xp 00000000 09:02 4560646 /usr/sbin/amcheck 00825000-00826000 rwxp 00008000 09:02 4560646 /usr/sbin/amcheck 008eb000-00904000 r-xp 00000000 09:02 2378079 /lib/ld-2.5.so 00904000-00905000 r-xp 00019000 09:02 2378079 /lib/ld-2.5.so 00905000-00906000 rwxp 0001a000 09:02 2378079 /lib/ld-2.5.so 0093a000-0093b000 r-xp 0093a000 00:00 0 [vdso] 0097b000-00ab5000 r-xp 00000000 09:02 2378081 /lib/libc-2.5.so 00ab5000-00ab7000 r-xp 0013a000 09:02 2378081 /lib/libc-2.5.so 00ab7000-00ab8000 rwxp 0013c000 09:02 2378081 /lib/libc-2.5.so 00ab8000-00abb000 rwxp 00ab8000 00:00 0 00ae7000-00aea000 r-xp 00000000 09:02 2378258 /lib/libtermcap.so.2.0.8 00aea000-00aeb000 rwxp 00002000 09:02 2378258 /lib/libtermcap.so.2.0.8 00dfb000-00e1b000 r-xp 00000000 09:02 4953736 /usr/lib/libamserver-2.5.0p2.so 00e1b000-00e1d000 rwxp 0001f000 09:02 4953736 /usr/lib/libamserver-2.5.0p2.so 00e1d000-00e1f000 rwxp 00e1d000 00:00 0 00ed0000-00eda000 r-xp 00000000 09:02 4674895 /usr/lib/libamtape-2.5.0p2.so 00eda000-00edb000 rwxp 00009000 09:02 4674895 /usr/lib/libamtape-2.5.0p2.so 00f9c000-00fbd000 r-xp 00000000 09:02 4675888 /usr/lib/libamanda-2.5.0p2.so 00fbd000-00fbe000 rwxp 00021000 09:02 4675888 /usr/lib/libamanda-2.5.0p2.so 00fbe000-00fdd000 rwxp 00fbe000 00:00 0 09fab000-09fcc000 rw-p 09fab000 00:00 0 b7e00000-b7e21000 rw-p b7e00000 00:00 0 b7e21000-b7f00000 ---p b7e21000 00:00 0 b7f86000-b7f89000 rw-p b7f86000 00:00 0 b7f92000-b7f94000 rw-p b7f92000 00:00 0 bfa86000-bfa9b000 rw-p bfa86000 00:00 0 [stack] Amanda Tape Server Host Check ----------------------------- Holding disk /share2/stage: 120193720 kB disk space available, using 120091320 kB
Nothing jumps out at me when I Google for the error. Tape drive and changer are:
[root@fraud etc]# cat /proc/scsi/scsi Attached devices: Host: scsi0 Channel: 00 Id: 04 Lun: 00 Vendor: HP Model: C1557A Rev: U709 Type: Sequential-Access ANSI SCSI revision: 02 Host: scsi0 Channel: 00 Id: 04 Lun: 01 Vendor: HP Model: C1557A Rev: U709 Type: Medium Changer ANSI SCSI revision: 02
Everything appears to work correctly otherwise (tape changer changes tapes, backups get done, etc.). Just amcheck and only when the next tape isn't in the changer is a problem.
Cheers, Dave