A couple days ago, Amanda started failing with the following errors in
the log that gets mailed to me:
FAILURE AND STRANGE DUMP SUMMARY:
zephyrus sda1 lev 0 FAILED [Request to zephyrus timed out.]
chinook sda5 lev 1 FAILED [could not connect to chinook]
chinook sda7 lev 0 FAILED [input: No such file or directory]
chinook sda7 lev 0 FAILED [dump to tape failed]
chinook sda6 lev 1 FAILED [dump to tape failed]
The first failure is OK (zephyrus is down right now), but I have no
idea why it couldn't connect to chinook. Chinook is the amanda
server. Here's the dump summary:
DUMPER STATS TAPER STATS
HOSTNAME DISK L ORIG-KB OUT-KB COMP% MMM:SS KB/s MMM:SS KB/s
-------------------------- --------------------------------------
--------------
chinook sda1 1 4659 1664 35.7 0:32 52.0 0:04 400.6
chinook sda5 1 FAILED --------------------------------------------
chinook sda6 1 FAILED --------------------------------------------
chinook sda7 0 FAILED --------------------------------------------
zephyrus sda1 0 FAILED --------------------------------------------
The strange thing is that sda1 (where the amanda configuraion files
are installed in /etc) gets backed up just fine.
Here's a snippet of what gets written in the full amdump log:
driver: adding holding disk 0 dir /space size 296960
reserving 296960 out of 296960 for degraded-mode dumps
driver: start time 225.611 inparallel 4 bandwidth 600 diskspace 296960 dir OBSOL
ETE datestamp 19990526 driver: drain-ends tapeq LFFO big-dumpers 1
driver: result time 225.611 from taper: TAPER-OK
driver: send-cmd time 225.611 to dumper0: FILE-DUMP 00-00001 /space/19990526/chi
nook.sda1.1 chinook sda1 1 1999:5:21:7:48:26 -2097087 DUMP |;bsd-auth;compress-f
ast;index;
driver: state time 225.611 free kps: 103 space: 275536 taper: idle idle-dumpers:
3 qlen tapeq: 0 runq: 3 stoppedq: 0 wakeup: 86400 driver-idle: no-bandwidth
driver: interface-state time 225.611 if : free 103
driver: hdisk-state time 225.611 hdisk 0: free 275536 dumpers 1
driver: result time 257.870 from dumper0: DONE 00-00001 4659 1664 32 [sec 32.016
kb 1664 kps 52.0 orig-kb 4659]
driver: finished-cmd time 257.971 dumper0 dumped chinook:sda1
driver: send-cmd time 257.971 to taper: FILE-WRITE 00-00002 /space/19990526/chin
ook.sda1.1 chinook sda1 1 19990526
driver: send-cmd time 257.971 to dumper0: FILE-DUMP 01-00003 /space/19990526/chi
nook.sda5.1 chinook sda5 1 1999:5:22:7:48:34 -2097087 DUMP |;bsd-auth;compress-f
ast;index;
driver: state time 257.971 free kps: 389 space: 238698 taper: writing idle-dumpe
rs: 3 qlen tapeq: 0 runq: 2 stoppedq: 0 wakeup: 86400 driver-idle: no-bandwidth
driver: interface-state time 257.971 if : free 389
driver: hdisk-state time 257.971 hdisk 0: free 238698 dumpers 1
driver: result time 258.024 from dumper0: TRY-AGAIN 01-00003 [could not connect
to data port: Resource temporarily unavailable]
Does anyone know what this "could not connect to data port" message
means? Amcheck doesn't produce any errors, so I don't know what to do
at this point. I recently installed an NIS server on the same
machine and it looks like the backups started failing around the same
time. Could this have something to do with it?
I'm using Debian potato with Linux kernel 2.2.3 and glibc 2.1.1, so
I'm cc-ing this to the Debian mailing list as well.
Please cc all replies to me directly.
Thanks,
Max
--
The hopeful depend on a world without end
Whatever the hopeless may say
Neil Peart, 1985
