On Sat, Aug 12, 2006 at 08:13:44PM +0100, Rodrigo Ventura wrote:
>
> Hello.
>
> I've recently upgraded to amanda 2.5.0p2 and switched the changer from the
> obscure chg-scsi to the much better chg-zd-mtx. All things have been running
> great for the past weeks, until yesterday.
>
> I got this error:
>
> *** A TAPE ERROR OCCURRED: [No writable valid tape found].
> Some dumps may have been left in the holding disk.
> Run amflush to flush them to tape.
> The next tape Amanda expects to use is: ISR003.
>
> This morning I did a "amtape ISR show" which resulted fine:
>
> changer: got exit: 0 str: 4 6 1
> changer_query: changer return was 6 1
> changer_query: searchable = 0
> amtape: scanning all 6 slots in tape-changer rack:
> changer_find: looking for NULL changer is searchable = 0
> changer: got exit: 0 str: 4 /dev/nst0
> slot 4: date 20060808 label ISR004
> changer: got exit: 0 str: 5 /dev/nst0
> slot 5: date 20060809 label ISR005
> changer: got exit: 0 str: 6 /dev/nst0
> slot 6: date 20060807 label ISR006
> changer: got exit: 0 str: 1 /dev/nst0
> slot 1: date 20060810 label ISR001
> changer: got exit: 0 str: 2 /dev/nst0
> slot 2: date 20060811 label ISR002
> changer: got exit: 0 str: 3 /dev/nst0
> slot 3: date 20060806 label ISR003
>
> Greping the amdump log one can find the error:
>
> changer: opening pipe to: /usr/local/amanda/libexec/chg-zd-mtx -info
> changer: got exit: 0 str: 5 6 1
> changer_query: changer return was 6 1
> changer_query: searchable = 0
> changer_find: looking for ISR003 changer is searchable = 0
> changer: opening pipe to: /usr/local/amanda/libexec/chg-zd-mtx -slot current
> changer: got exit: 0 str: 5 /dev/nst0
> changer: opening pipe to: /usr/local/amanda/libexec/chg-zd-mtx -slot next
> changer: got exit: 0 str: 6 /dev/nst0
> changer: opening pipe to: /usr/local/amanda/libexec/chg-zd-mtx -slot next
> changer: got exit: 0 str: 1 /dev/nst0
> changer: opening pipe to: /usr/local/amanda/libexec/chg-zd-mtx -slot next
> changer: got exit: 0 str: 2 /dev/nst0
> changer: opening pipe to: /usr/local/amanda/libexec/chg-zd-mtx -slot next
> changer: got exit: 0 str: 3 /dev/nst0
> changer: opening pipe to: /usr/local/amanda/libexec/chg-zd-mtx -slot next
> changer: got exit: 0 str: 4 /dev/nst0
> slot 5: read label `ISR005', date `20060809'
> slot 6: read label `ISR006', date `20060807'
> slot 1: read label `ISR001', date `20060810'
> slot 2: read label `ISR002', date `20060811'
> slot 3: not an amanda tape (Input/output error)
> slot 4: read label `ISR004', date `20060808'
>
> driver: result time 491.656 from taper: TAPE-ERROR [slot 5: read label
> `ISR005', date `20060809'
> slot 6: read label `ISR006', date `20060807'
> slot 1: read label `ISR001', date `20060810'
> slot 2: read label `ISR002', date `20060811'
> slot 3: not an amanda tape (Input/output error)
> slot 4: read label `ISR004', date `20060808'
>
> So it seems the tape on slot 3 has an error. However, I ran amverify on tape
> ISR003 and no errors were found!
>
> $ amtape ISR slot 3
> changer: got exit: 0 str: 3 /dev/nst0
> amtape: changed to slot 3 on /dev/nst0
> $ amtape ISR current
> changer: got exit: 0 str: 3 6 1
> changer_query: changer return was 6 1
> changer_query: searchable = 0
> amtape: scanning current slot in tape-changer rack:
> changer: got exit: 0 str: 3 /dev/nst0
> slot 3: date 20060806 label ISR003
>
> Any clues?
>
My HP 6 slot DDS changer has the nasty habit of
returning to the OS as if it has completed a task,
like rewind, change slots, eject, load magazine, etc.
Then as far as the OS and software are concerned,
the next command, like read the header, can be done.
This command would sometimes fail because the tape
was still rewinding or changing or ???.
I put a couple of sleep's in the chg-mtx script
(I wasn't using chg-zd-mtx). Got to the point
that I could live with the delays caused by several
sleeps by starting them out long, then backing off
til the problems reappeared.
--
Jon H. LaBadie jon AT jgcomp DOT com
JG Computing
4455 Province Line Road (609) 252-0159
Princeton, NJ 08540-4322 (609) 683-7220 (fax)
|
