Problem finishing old backup .. !?
This morning, I cam back from holiday, and had to see that there was no successful backup since some days ... :-( amstatus shows me, that almost all hosts have been backed up, but some still have dumping to tape running for over 26 hours, although there is no running amanda-process or anything else ... I suppose, that amandas logfiles are in an inconsistent state !? I can run amcheck without any error ... ... 0 dumpers busy : 0:00:00 ( 0.00%) 1 dumper busy : 5:49:25 ( 9.80%)not-idle: 5:49:25 (100.00%) 2 dumpers busy : 0:00:30 ( 0.01%)not-idle: 0:00:30 (100.00%) 3 dumpers busy : 0:00:30 ( 0.01%)not-idle: 0:00:30 (100.00%) 4 dumpers busy : 0:00:30 ( 0.01%)not-idle: 0:00:30 (100.00%) 5 dumpers busy : 0:00:30 ( 0.01%)not-idle: 0:00:30 (100.00%) 6 dumpers busy : 0:00:30 ( 0.01%)not-idle: 0:00:30 (100.00%) 7 dumpers busy : 0:00:30 ( 0.01%)not-idle: 0:00:30 (100.00%) 8 dumpers busy : 0:00:30 ( 0.01%)not-idle: 0:00:30 (100.00%) 9 dumpers busy : 0:00:30 ( 0.01%)not-idle: 0:00:30 (100.00%) 10 dumpers busy : 0:00:30 ( 0.01%)not-idle: 0:00:30 (100.00%) 11 dumpers busy : 0:00:30 ( 0.01%)not-idle: 0:00:30 (100.00%) 12 dumpers busy : 0:00:30 ( 0.01%)not-idle: 0:00:30 (100.00%) 13 dumpers busy : 0:00:30 ( 0.01%)not-idle: 0:00:30 (100.00%) 14 dumpers busy : 0:00:30 ( 0.01%)not-idle: 0:00:30 (100.00%) 15 dumpers busy : 0:00:30 ( 0.01%)not-idle: 0:00:30 (100.00%) 16 dumpers busy : 0:00:30 ( 0.01%)not-idle: 0:00:30 (100.00%) 17 dumpers busy : 0:00:30 ( 0.01%)not-idle: 0:00:30 (100.00%) 18 dumpers busy : 0:00:30 ( 0.01%)not-idle: 0:00:30 (100.00%) 19 dumpers busy : 0:00:30 ( 0.01%)not-idle: 0:00:30 (100.00%) 20 dumpers busy : 0:00:30 ( 0.01%)not-idle: 0:00:30 (100.00%) 21 dumpers busy : 0:00:30 ( 0.01%)not-idle: 0:00:30 (100.00%) 22 dumpers busy : 0:00:30 ( 0.01%)not-idle: 0:00:30 (100.00%) 23 dumpers busy : 0:00:30 ( 0.01%)not-idle: 0:00:30 (100.00%) 24 dumpers busy : 0:00:30 ( 0.01%)not-idle: 0:00:30 (100.00%) 25 dumpers busy : 0:00:30 ( 0.01%)not-idle: 0:00:30 (100.00%) 26 dumpers busy : 0:00:30 ( 0.01%)not-idle: 0:00:30 (100.00%) 27 dumpers busy : 0:00:30 ( 0.01%)not-idle: 0:00:30 (100.00%) 28 dumpers busy : 0:00:30 ( 0.01%)not-idle: 0:00:30 (100.00%) 29 dumpers busy : 0:00:30 ( 0.01%)not-idle: 0:00:30 (100.00%) 30 dumpers busy : 0:00:30 ( 0.01%)not-idle: 0:00:30 (100.00%) 31 dumpers busy : 0:00:30 ( 0.01%)not-idle: 0:00:30 (100.00%) 32 dumpers busy : 0:00:30 ( 0.01%)not-idle: 0:00:30 (100.00%) 33 dumpers busy : 0:00:30 ( 0.01%)not-idle: 0:00:30 (100.00%) 34 dumpers busy : 0:00:30 ( 0.01%)not-idle: 0:00:30 (100.00%) 35 dumpers busy : 0:00:30 ( 0.01%)not-idle: 0:00:30 (100.00%) 36 dumpers busy : 0:00:30 ( 0.01%)not-idle: 0:00:30 (100.00%) 37 dumpers busy : 0:00:30 ( 0.01%)not-idle: 0:00:30 (100.00%) 38 dumpers busy : 0:00:30 ( 0.01%)not-idle: 0:00:30 (100.00%) 39 dumpers busy : 0:00:30 ( 0.01%)not-idle: 0:00:30 (100.00%) 40 dumpers busy : 0:00:30 ( 0.01%)not-idle: 0:00:30 (100.00%) 41 dumpers busy : 0:00:30 ( 0.01%)not-idle: 0:00:30 (100.00%) 42 dumpers busy : 0:00:30 ( 0.01%)not-idle: 0:00:30 (100.00%) Were great, anybody had a hint ... ...olli
amanda + virtual interfaces
Hi everyone, Has anyone of you experiences or does anyone know if it is possible to let amanda run over a virtual interface ? We are trying to implement a new backupserver, which shall replace the old one after some weeks of parrallel running. The problem is, the we have to put the new server in 46 hosts´ .amandahosts and don´t want to do the same again in some weeks. So we thought about giving the new server a virtual ip, attach a dnsname like backup1 to it, and put that in all .amandahosts BigPic: - lemmy -- current hot backupservers dns-name - neo-lemmy -- current new hot backupservers dns-name - neo-lemmy2 -- current cold-standby backup, will become hot in the next weeks. -- - lemmy is in all .amandahosts - neo-lemmy has to be put in if lemmy goes offline, name neo-lemmy will be switched to name lemmy the aim would be, put -for example- backup1 in all .amandahosts. Give neo-lemmy a virtual-ip, and attach the name backup1 to that ip. In this way, we could switch the primary hostname to lucifer or whatever, and had no problem with the .amandahosts files on the clients. Thanks in advance, ...olli
sporadic errors - could not connect ...
Hi List ! Were great anyone could help ... We are experiencing strange errors at the time. We have an amand-server that backs up about 30 hosts a night. Most of the time it went well, but now there are very often etrange errors. It sais: x /opt lev 0 FAILED 20040606[could not connect to x] And that on about 15 of 30 hosts. The more interesting is, that the first filesystems (or dle´s) are fine, only some of them are failing. Next day everything is fine, or other filesystems on other hosts are failing .. !? On some days There seem to be no networking problems, all hosts are 100 Mbit connected by a 1 Gbit HP Switch separated in VLANS. The amanda-server is a 1GHz Pentium 3 with 2 2TB RAids attached to it, so Space should not be the problem. I wrote a little script, that monitors wih amdump the backup every 3 minutes in a file named date-time, but the output doesnt´t give any hints ... Neither the amanda-server, nor the clients tell anything in the log-files ... !? Would it be an idea to set (if existent) a higher debug-level at the amanda-server, and where should i do that ? Thanks in Advance ... ...olli
amandad Problem on redhat 7.2
Hi List-Members ...
At the moment we are experiencing another problem on one of our clients.
Were great anyone had a hint for us to solve this.
After starting xinetd on the client ---
/etc/xinetd.d/amandad
# default: off
# description: Amanda backup client
#
service amanda
{
socket_type = dgram
protocol= udp
wait= no
user= amanda
group = disk
server = /usr/lib/amanda/amandad
disable = no
}
/
(# ll /usr/lib/amanda/amandad
-rwxr-xr-x1 rpm disk42617 Sep 12 2002
/usr/lib/amanda/amandad )
--- everything looks fine. If we start an amcheck DailySet1 the only
client that fails is this redhat machine.
In /var/log/messages you can see the following:
Apr 30 11:47:52 god xinetd[7397]: Started working: 3 available services
Apr 30 11:47:54 god xinetd: Starten von xinetd succeeded
Apr 30 11:47:59 god xinetd[7397]: file descriptor of service amanda has
been closed
Apr 30 11:47:59 god xinetd[7397]: select reported EBADF but no bad file
descriptors were found
Machine Data is:
-
Linux version 2.4.7-10smp ([EMAIL PROTECTED]) (gcc
version 2.96 2731 (Red Hat Linux 7.1 2.96-98))
-
build: VERSION=Amanda-2.4.2p2
BUILT_DATE=Thu Sep 12 18:06:07 UTC 2002
BUILT_MACH=Linux D5 2.4.19 #1 SMP Tue Sep 10 00:06:19 UTC 2002
i686 unknown
CC=gcc
-
xinetd from rpm: xinetd-2.3.11-1.7x
-
Can anyone help ?
Thanks in advance,
...olli
Re: amandad Problem on redhat 7.2
Hi group !
Thanks for your advices !
I tried Pauls suggestion an corrected the ownership-problem Bernard
found ...
All fine now ! Thanks a lot !
...olli
Oliver Simon wrote:
Hi List-Members ...
At the moment we are experiencing another problem on one of our clients.
Were great anyone had a hint for us to solve this.
After starting xinetd on the client ---
/etc/xinetd.d/amandad
# default: off
# description: Amanda backup client
#
service amanda
{
socket_type = dgram
protocol= udp
wait= no
user= amanda
group = disk
server = /usr/lib/amanda/amandad
disable = no
}
/
(# ll /usr/lib/amanda/amandad
-rwxr-xr-x1 rpm disk42617 Sep 12 2002
/usr/lib/amanda/amandad )
--- everything looks fine. If we start an amcheck DailySet1 the only
client that fails is this redhat machine.
In /var/log/messages you can see the following:
Apr 30 11:47:52 god xinetd[7397]: Started working: 3 available services
Apr 30 11:47:54 god xinetd: Starten von xinetd succeeded
Apr 30 11:47:59 god xinetd[7397]: file descriptor of service amanda has
been closed
Apr 30 11:47:59 god xinetd[7397]: select reported EBADF but no bad file
descriptors were found
Machine Data is:
-
Linux version 2.4.7-10smp ([EMAIL PROTECTED]) (gcc
version 2.96 2731 (Red Hat Linux 7.1 2.96-98))
-
build: VERSION=Amanda-2.4.2p2
BUILT_DATE=Thu Sep 12 18:06:07 UTC 2002
BUILT_MACH=Linux D5 2.4.19 #1 SMP Tue Sep 10 00:06:19 UTC 2002
i686 unknown
CC=gcc
-
xinetd from rpm: xinetd-2.3.11-1.7x
-
Can anyone help ?
Thanks in advance,
...olli
Problems while amanda is running !?
Hi group ! Maybe anyone can help me solving this problem or had the same probs ? After running amanda for quite long time without any problems, the following messages crop up in my mails every morning. Can anyone help ? The only thing we changed, was to put a new switch online, that has GBit-Capabilities. The mail every morning really looks shi$, because of 272 DLE´s and not a half of them can be done without problems. -snip hal/var lev 1 FAILED 20040420[could not connect to hal] hydra /opt lev 0 FAILED 20040420[could not connect to hydra] helena1/boot lev 0 FAILED 20040420[could not connect to helena1] -/snip Were great, anyone had a hint ! Greetings, ...oliver
Amplot - ps-File ?
Hi Group ! It´s just a nice to have, but maybe anyoune can help .. ? Does anyone know a possibility to put amplot-stats in ps-Files to move them to jpg´s or something like that ? Amplot has only the ablility to push the graphics to display (?), ps-files were much better for our needs to organize them in a directory structure and view them as html-embedded jpgs. Looking forward to your answers ...olli
Urgent help needed, degraded mode, tapes too small ?
Hi Group ! This smorning we got the problem, that many servers were backuped in degraded mode (host /disk lev 1 FAILED). I think this means, the defined hdd-tapes became too small, right ? 1) We have 17 tapes defined each one with 75000 MB. Can I expand this to 10 MB without loosing the current backups ? 2) What other options do I have ? Last night, the backups partial ran in degraded mode, will amanda try again to do those dumps additional to the normal run, will it continue normally, or what ? 3) What if I remove some entries from the disklist ? Will amanda recognize the removed parts although these could not be backuped fine last night ? Some of the data could be removed from the disklist, until we get the new backup-servers and raids to backup to, but until that, we still need to do a backup .. !? Backup needs about 4 1/2 hours by now, if amanda would also try to do something from last night, we will run into some problems ... Thanks in advance, ...olli
Re: Urgent help needed, degraded mode, tapes too small ?
Hi again ... Am Di, 2004-01-27 um 16.43 schrieb Jon LaBadie: On Tue, Jan 27, 2004 at 12:32:06PM +0100, Oliver Simon wrote: Hi Group ! This smorning we got the problem, that many servers were backuped in degraded mode (host /disk lev 1 FAILED). I think this means, the defined hdd-tapes became too small, right ? 1) We have 17 tapes defined each one with 75000 MB. Can I expand this to 10 MB without loosing the current backups ? Stefan provided answers, I'm curious though. Each of your vtapes is 75GB. From your past experience in the amdump reports, how full have they been? If they have regularly been half or more, say 40GB/amdump run, does your total disk storage allocated for vtapes support this? I.e. it will take perhaps 17*40GB or 680GB total for your collection of vtapes. Of course now that you are filling 75GB vtapes, and considering making them 100GB, will you be allocating a terrabyte of disk storage to your vtapes? FilesystemSize Used Avail Use% Mounted on /dev/sdb1 1.3T 1006G 254G 80% /r01 /dev/sdc1 2.8T 611G 2.2T 24% /r02 --- DailySet1 /dev/sdd1 2.0T 1.3T 712G 65% /r03 I guess what I'm driving at, is your problem that 75GB per vtape is too small or is the total storage capacity too small. Did the file system run out of space? Nope, I don´t think so ...
disk offline ?
Hi Group ! Maybe anyone has a hint for me ? We have about 30 Servers, all UNIX. They are all backed up with amanda, which worked very good, until some problems occured. I defined a TestSet an tried to put in a new host. But I don´t get it to work. I want to do a dump of the following partitions, which are really partitions on the target-host. -snip disklist daisy / comp-hd-dump daisy /boot comp-hd-dump daisy /optcomp-hd-dump daisy /usrcomp-hd-dump -/snip If I put comp-hd-tar in here, all works fine ... !? -snip report FAILURE AND STRANGE DUMP SUMMARY: daisy /usr lev 0 FAILED [disk /usr offline on daisy?] daisy /opt lev 0 FAILED [disk /opt offline on daisy?] daisy /lev 0 FAILED [disk / offline on daisy?] ... DUMP SUMMARY: DUMPER STATSTAPER STATS HOSTNAME DISK L ORIG-KB OUT-KB COMP% MMM:SS KB/s MMM:SS KB/s -- - daisy / 0 FAILED --- daisy /boot 06410 5472 85.4 0:023511.7 0:023484.9 daisy /opt0 FAILED --- daisy /usr0 FAILED --- /snip Why does /boot work ??? Ups, just saw it worked only the first time ... Now its the same message ... And, does comp-hd-tar eat more cpu ? I would think so, or am I false ? Last Backup this night Were great anyone had a tip ! ...olli
Re: disk offline ?
I can su to amanda and chdir to usr and everything else thats failling ? Is it maybe because of the filesystem xfs on the target-machine ? I think amanda wants to do a xfsdump there, right ? I tried to run it as amanda and got it complaining to be root to run ... !? Thanks in advance ...olli Am Do, 2004-01-08 um 16.39 schrieb Frank Smith: --On Thursday, January 08, 2004 15:51:21 +0100 Oliver Simon [EMAIL PROTECTED] wrote: Hi Group ! Maybe anyone has a hint for me ? We have about 30 Servers, all UNIX. They are all backed up with amanda, which worked very good, until some problems occured. I defined a TestSet an tried to put in a new host. But I don´t get it to work. I want to do a dump of the following partitions, which are really partitions on the target-host. -snip disklist daisy / comp-hd-dump daisy /boot comp-hd-dump daisy /optcomp-hd-dump daisy /usrcomp-hd-dump -/snip If I put comp-hd-tar in here, all works fine ... !? -snip report FAILURE AND STRANGE DUMP SUMMARY: daisy /usr lev 0 FAILED [disk /usr offline on daisy?] daisy /opt lev 0 FAILED [disk /opt offline on daisy?] daisy /lev 0 FAILED [disk / offline on daisy?] Disk offline messages are generally a result of permissions problems (if the disk were really offline you would have much bigger problems). Make sure your backup user is part of the correct group to have access to the disk. Try to 'su - backupuser' and check if you can see those disks. Tar works because Aamnda uses a suid runtar wrapper script that runs tar as root (although it isn't used by amcheck, so you would still get errors from amcheck even though amdump would work). Frank ... DUMP SUMMARY: DUMPER STATSTAPER STATS HOSTNAME DISK L ORIG-KB OUT-KB COMP% MMM:SS KB/s MMM:SS KB/s -- - daisy / 0 FAILED --- daisy /boot 06410 5472 85.4 0:023511.7 0:023484.9 daisy /opt0 FAILED --- daisy /usr0 FAILED --- /snip Why does /boot work ??? Ups, just saw it worked only the first time ... Now its the same message ... And, does comp-hd-tar eat more cpu ? I would think so, or am I false ? Last Backup this night Were great anyone had a tip ! ...olli
Re: disk offline ?
Hello again ... Am Do, 2004-01-08 um 17.37 schrieb Frank Smith: --On Thursday, January 08, 2004 17:21:48 +0100 Oliver Simon [EMAIL PROTECTED] wrote: I can su to amanda and chdir to usr and everything else thats failling ? Is it maybe because of the filesystem xfs on the target-machine ? I think amanda wants to do a xfsdump there, right ? I tried to run it as amanda and got it complaining to be root to run ... !? You must use the right version of dump for the filesystem. Was Amanda built with support for xfsdump ? Yes, we have about 15 (SuSI´s/SLES/SLOX) hosts running with xfs. No problem there. Are the disks on all your clients xfs or just on the one new client? The rest is 10 Solaris (ufs) and some other linux´ (redhat ext3). If you have a mixture of filesystem types, I'm not sure how you configure multiple flavors of dump in one setup, hopefully someone else here has done it and will say how. Thought, amanda examines it itself, and takes the correct version (dump/xfsdump) from the clients ?? Frank ...olli Thanks in advance ...olli Am Do, 2004-01-08 um 16.39 schrieb Frank Smith: --On Thursday, January 08, 2004 15:51:21 +0100 Oliver Simon [EMAIL PROTECTED] wrote: Hi Group ! Maybe anyone has a hint for me ? We have about 30 Servers, all UNIX. They are all backed up with amanda, which worked very good, until some problems occured. I defined a TestSet an tried to put in a new host. But I don´t get it to work. I want to do a dump of the following partitions, which are really partitions on the target-host. -snip disklist daisy / comp-hd-dump daisy /boot comp-hd-dump daisy /optcomp-hd-dump daisy /usrcomp-hd-dump -/snip If I put comp-hd-tar in here, all works fine ... !? -snip report FAILURE AND STRANGE DUMP SUMMARY: daisy /usr lev 0 FAILED [disk /usr offline on daisy?] daisy /opt lev 0 FAILED [disk /opt offline on daisy?] daisy /lev 0 FAILED [disk / offline on daisy?] Disk offline messages are generally a result of permissions problems (if the disk were really offline you would have much bigger problems). Make sure your backup user is part of the correct group to have access to the disk. Try to 'su - backupuser' and check if you can see those disks. Tar works because Aamnda uses a suid runtar wrapper script that runs tar as root (although it isn't used by amcheck, so you would still get errors from amcheck even though amdump would work). Frank ... DUMP SUMMARY: DUMPER STATSTAPER STATS HOSTNAME DISK L ORIG-KB OUT-KB COMP% MMM:SS KB/s MMM:SS KB/s -- - daisy / 0 FAILED --- daisy /boot 06410 5472 85.4 0:023511.7 0:023484.9 daisy /opt0 FAILED --- daisy /usr0 FAILED --- /snip Why does /boot work ??? Ups, just saw it worked only the first time ... Now its the same message ... And, does comp-hd-tar eat more cpu ? I would think so, or am I false ? Last Backup this night Were great anyone had a tip ! ...olli
