RE: client timing out
Title: client timing out Hi, the amandad on the client host ist probably not running. According to the system you use you have to enter something like the following: /etc/services: amanda 10080/udp amandaidx 10082/tcp amidxtape 10083/tcp /etc/inetd.conf: amanda dgram udp wait amanda /usr/local/libexec/amandad amandad After this, don't forgett to send a kill -1 to the pid of the inetd. Wolfgang -Original Message-From: Anthony Carter [mailto:[EMAIL PROTECTED]]Sent: Mittwoch, 16. Mai 2001 17:28To: '[EMAIL PROTECTED]'Subject: client timing out I installed Amanda and I'm trying to backup the client (also the Amanda host). I am getting the message that the client is timing out. Any suggesstions? Thanks Anthony Carter
Re: Problems with amcheck and hostname resolution.
>I recently decided to add another Amanda client to my existing Amanda
>setup (server running v2.4.1p1) where I back up three clients using Krb4.
Does the new client also use Krb4?
>ERROR: shell1.nominum.com: couldn't resolve hostname
Here's the relevant code:
#ifdef KRB4_SECURITY
if(hostp->disks->auth == AUTH_KRB4)
rc = make_krb_request(hostp->hostname, kamanda_port, req,
hostp, CHECK_TIMEOUT, handle_response);
else
#endif
rc = make_request(hostp->hostname, amanda_port, req,
hostp, CHECK_TIMEOUT, handle_response);
req = NULL; /* do not own this any more */
if(rc) {
/* couldn't resolve hostname */
fprintf(outf,
"ERROR: %s: couldn't resolve hostname\n", hostp->hostname);
So what the message really means is that make_krb_request (if you're
using krb4) or make_request sent back a non-zero return code.
It's true that a simple make_request call can only return non-zero if
gethostbyname fails.
However, make_krb_request can fail if either host2krbname fails or if
add_krb_security fails. Host2krbname can only fail if gethostbyname
fails, so we're back to the ability of the Amanda server to be able to
look up shell1.nominum.com (see below).
Add_krb_security calls get_krb_security which can fail if krb_mk_req
(which appears to be a low level K4 routine?) fails. For that it's
working with CLIENT_HOST_PRINCIPLE, CLIENT_HOST_INSTANCE, the realm and
the incoming request.
The first thing I'd do is make sure gethostbyname works for that client.
I'd do that with:
ftp://gandalf.cc.purdue.edu/pub/amanda/gethostbyname.c
You cannot trust nslookup to give you the right answer -- it only tests
DNS, not the other possible host name sources (e.g. /etc/hosts).
If none of that helps, I'd start stepping through the K4 code and see what
it didn't like and (hopefully) why. But K4 is only minimally supported in
Amanda, and I certainly don't know much about it other than the basics.
I'm just hoping the above overview will jog some memory of what you did
on one of the other clients that does work.
Clearly that message should be cleaned up a bit. I've added that to the
TODO list.
>Thanks - Peter
John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
Re: oh oh...
>Amanda started returning these results a couple of days ago... U, let's be clear here. Amanda is not saying this. Your dump program is. Amanda is just passing along the good news :-). >? DUMP: bread: lseek fails This might be nothing more than an active file system. What happens is dump reads what it thinks are going to be disk block addresses (e.g. from an indirect block) and by the time it gets there, that disk block has been re-assigned to some file for data, so when it tries to seek based on what it thinks is a disk block, it jumps off into never never land. As long as the dump goes ahead and completes, I wouldn't worry about it. Or you could switch to GNU tar (but check the mailing list for the issues that brings up). You might also make sure you're up to date on dump version. A lot of good things have happened to it in the last several months (so I've heard). >Anything I can check for? Is there anyway I can identify where on the disk >it's failing, and try removing that file? You could get the source code for dump and add some better diagnostics (what you're getting is, shall we say, marginal). That's probably not too hard -- just some extra print statements. >Chris Herrmann John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
oh oh...
Hi, Amanda started returning these results a couple of days ago... FAILED AND STRANGE DUMP DETAILS: /-- gemini sda5 lev 0 FAILED [/sbin/dump returned 3] sendbackup: start [gemini:sda5 level 0] sendbackup: info BACKUP=/sbin/dump sendbackup: info RECOVER_CMD=/bin/gzip -dc |/sbin/restore -f... - sendbackup: info COMPRESS_SUFFIX=.gz sendbackup: info end | DUMP: Date of this level 0 dump: Thu May 17 03:01:30 2001 | DUMP: Date of last level 0 dump: the epoch | DUMP: Dumping /dev/sda5 (/home) to standard output | DUMP: Label: none | DUMP: mapping (Pass I) [regular files] | DUMP: mapping (Pass II) [directories] | DUMP: estimated 3210845 tape blocks. | DUMP: Volume 1 started at: Thu May 17 03:01:45 2001 | DUMP: dumping (Pass III) [directories] | DUMP: dumping (Pass IV) [regular files] | DUMP: 4.66% done, finished in 1:42 | DUMP: 12.08% done, finished in 1:12 | DUMP: 20.04% done, finished in 0:59 ? DUMP: bread: lseek fails ... So I took the server down to runlevel 1, e2fsck'd the drive checking for physical bad blocks as well as damaged inodes etc. It said that it found a problem with 1 file, which it repaired. Brought it back up, and the next night had the same problem. Thought it might have been related to 1 very large, recently saved file (1.3G), but made it much smaller (400M) and it's still happening. Anything I can check for? Is there anyway I can identify where on the disk it's failing, and try removing that file? Thanks, Chris Herrmann Far Edge Technology p. 02 99553640 f. 02 99547994 m. 0403 393309 http://www.faredge.com.au
Problems with amcheck and hostname resolution.
I recently decided to add another Amanda client to my existing Amanda setup (server running v2.4.1p1) where I back up three clients using Krb4. I went through the usual process in compiling & installing the client binaries, but doing the host check on the Amanda server end comes up with this: -=- Amanda Backup Client Hosts Check ERROR: shell1.nominum.com: couldn't resolve hostname Client check: 4 hosts checked in 0.320 seconds, 1 problem found. -=- Granted, it has been a couple of years since I have had to add a new Amanda client, so I might be missing something completely obvious. The server name in question meets all the lookup requirements (forward and reverse IP lookups) The other clients Amanda backs up are coming back with no problems, so I am rather stumped at what the problem could be with this new client. Anyone have any ideas? Thanks - Peter -- [EMAIL PROTECTED] - [ Systems Admin. | Nominum, Inc. ]
Re: full back up
>All was working well, but since a few weeks my weekly configuration is >trying to do a level 2 backup of the disk /usr2. ... Others are reporting problems with "strategy noinc". I'd need to see the amdump. file that went along with Amanda doing something other than a level 0. >Last week i forced a full backup of >this disk via >amadmin, but the last backup failed again with > >FAILURE AND STRANGE DUMP SUMMARY: > gretel /usr2 lev 0 FAILED [can't switch to incremental dump] Again, I'd need to see the amdump. file. As a guess, you might be out of holding disk space. That would cause Amanda to fail to do the full dump and also not switch to an incremental (either because of the "force" or "strategy noinc"). Or it could be something else. It might also be useful to grep for /usr2 in the log.. file and see whether any other messages were present, but amreport just didn't display them. >I have nothing changed on the Amanda server, in the configurations or >something else. So why is Amanda trying to do incrementals since a >few weeks? What makes you think it's trying to do incrementals? I didn't see any evidence of that in the report you sent. The "can't switch to incremental dump" message doesn't mean it tried to do an incremental. It means it couldn't do a full and while it normally would have tried an incremental, something prevented that. I agree there should be more messages about why it failed. >Jens Krause John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED] P.S. You might take a look at columnspec in amanda.conf. It would clean up the Amanda reports so the columns don't run together.
Re: backups started failing
On Wed, 16 May 2001, John R. Jackson wrote: > >I've been running Amanda 2.5.0 (out of CVS) for about 4 months now. ... > > You do understand that's the development branch, right? In other words, > you should consider yourself lucky if it even compiles, let alone runs, > let alone "does the right thing". yes. It seems to be working fairly well so far. I haven't had any problems except for this. > > > paddedcell /export/home lev 0 FAILED [Request to > >paddedcell.psychoses.org failed: timeout waiting for REP] > >... > >Looking at the /tmp/*debug files, I don't see anything obvious, except for > >"amandad: received dup P_REQ packet, ACKing it" in amanda.debug. > > Does amandad say it sent back the REP? I didn't see anything about that. The problem started rougly about 3 days after I changed the filesystems on this machine to use "compress client fast". I moved it back to no compression and paddedcell now backs up fine again. I'll see if I can chase things down. > >Ideas? > > Start leaning C and debugging :-). So I have to start bringing work home with me? *grin*
Re: backups started failing
>I've been running Amanda 2.5.0 (out of CVS) for about 4 months now. ... You do understand that's the development branch, right? In other words, you should consider yourself lucky if it even compiles, let alone runs, let alone "does the right thing". > paddedcell /export/home lev 0 FAILED [Request to >paddedcell.psychoses.org failed: timeout waiting for REP] >... >Looking at the /tmp/*debug files, I don't see anything obvious, except for >"amandad: received dup P_REQ packet, ACKing it" in amanda.debug. Does amandad say it sent back the REP? >Ideas? Start leaning C and debugging :-). John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
Re: Backup in one tape
>-i don't manage to make incremental backup of many >disk :( >into one tape (it seems to make a full backup each >amdump) That's correct. Every time you run Amanda (amdump or amflush), it will start a new tape. That's just the way it is (for now). If you want to conserve tapes, and you are sure you understand the danger of leaving things in the holding disk without a copy on tape, you can just not load a tape in the drive and amdump will leave things in the holding disk. When you have enough there to make it worth while, run amflush and write it all out to tape. >Actually i suppose when amanda reuse a tape, it >simply erase it, amanda does this if tape contains >other disks for an another configuration (ex: >DailySet2)??? When Amanda reuses a tape, it just starts writing again at the beginning. If you have multiple configurations, you should use labelstr and amlabel to make sure each configuration has a unique label. This will protect you from accidentally using the wrong tape in the wrong configuration. For instance, our Research Computing Division (RCD) has a configuration and the Instructional Computing Division (ICD) has a different one. The tape labels are "nn/RCD" and "nn/ICD" to make sure they cannot get mixed up. >Beside, i reaffirm a little bug with amrecover: >>EOF, check amidxtaped.debug file on localhost. >>... >and little : 'mt rewind' is adviced before OR retry a >2nd extract with amrecover -> that's works well. Could you explain this some more? Normally amidxtaped (used by amrecover) rewinds the tape after each request. I don't know why you would need to explicitly do it. >I didn't found into amanda commands for deleting a >disk into a tapes(not into the database definitely), >is it possible with amanda or other command??? Not sure I understand. If you want Amanda to forget about a particular disk that has been backed up before, you can use the "delete" subcommand of "amadmin" to remove it from the database. John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
Amlabel and /dev/nht0
Amanda was working so well in the server room that I decided to slap it on my workstation as well. I have a Seagate STT2A drive in my linux box. It works just fine with Tar and the couple of times I tried out an eval version of BRU. However, when I run amlabel to label my tapes, it says that the tape is not an amanda tape, which is fine as it's brand new, then it writes a label, but when it goes to check it afterwards it says "No label found" and "Are you sure /dev/nht0 is a non rewinding device?" mt -f /dev/nht0 rewind and dd if=/dev/nht0 bs=32k count=1 yield nothing. I checked to see if it was indeed a non-rewinding device by doing successive dd commands on a previously tar'd tape and it advanced along the tape without rewinding just as I would expect it to. I'm using the tapetype definition from the Faq-o-matic for the STT2A. Any ideas on what would cause this?
Re: Lev 0 backup to disk - single system
>Does anyone know how to get amanda to run a full backup but leave the >dumpfiles on the holding disk after writing to tape? ... This comes up here every once in a while, and eventually we'll get some real support for it in the code. One recent (around the first of this month) discussion had this Subject: How to tell Amanda to keep the latest full+incremental on the holding disk? Check the E-mail archives. It includes a source hack that will relink the files into a parallel area from the holding disk as they are being removed. >David Carter John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
Re: client timing out
>I installed Amanda and I'm trying to backup the client (also the Amanda >host). I am getting the message that the client is timing out. Any >suggesstions? ... The best suggestion is to go to the FAQ at www.amanda.org. This error is very common for new installations and is covered in detail there. If you still cannot get it going, please post back with what version of Amanda you're using, what OS, what steps you tried and what the results were. As I said, this is a very common setup problem, so I'm sure we can help you get it going. >Anthony Carter John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED] P.S. Please turn off "send as HTML" in your mailer. It's just a waste of bandwidth.
Re: Device ..
>How would I direct Amanda to make a backup on a disk on the system?
I'm not sure I understand the question. Do you want to do your backups
but leave them on disk instead of writing to tape? If so, there are a
couple of possibilities. One is to just not mount a tape and let Amanda
leave things in the holding disk. There are some issues with this, such
as having to set the "reserve" parameter so it will do full backups and
I'm not sure it will bump to anything other than a level 1.
You could also get the amanda-242-tapeio branch from CVS (see the FAQ
for details). It is under development but some people are using it.
There is support for alternate output devices, one of which ("file")
uses a disk directory as a pseudo tape.
>Dr Prabhakar Ganapathy ( Dr.Prab)
John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
P.S. Please turn off "send as HTML" in your mail program. It's just
a waste of bandwidth.
Backup in one tape
Hello, First, I manage to install and work Amanda in a redhat 7.1 (kernel2.4.3) with an ADR DI-30 with scsi emulation(osst) and samba(v2.0.7) Thank for all for many informations that i found in the mailing list archive :)) --here is a lot of questions about the use of amanda -i don't manage to make incremental backup of many disk :( into one tape (it seems to make a full backup each amdump) this my paramater of amanda.conf: >dumpcycle 1 weeks >runspercycle 7 days >tapecycle 1 tapes this is the disklist: leto /home/toto comp-test-recover cable /var/lib/amanda comp-user-tar-recover cable /root comp-user-tar-recover cable /libcomp-user-tar-recover But none files was modified between two runs, so incrementation level 1 is stupid or i understand nothing with incremental method. --amreport of 24/05: HOSTNAME DISKL ORIG-KB OUT-KB COMP% MMM:SS KB/s MMM:SS KB/s -- - cable/lib1 37290 14112 37.8 0:37 381.4 0:26 537.8 cable/root 0 320288 90.0 0:01 420.3 0:03 120.0 cable-lib/amanda 0 40 32 80.0 0:00 181.5 0:02 28.6 leto /home/toto 0 20 32 160.0 0:01 43.3 0:07 9.6 with level(amoverview): host disk ...05 24 cable/lib 1 cable/root 1 cable/var/lib/amanda1 leto /home/toto 1 --amreport of 25/05: DUMP SUMMARY: DUMPER STATS TAPER STATS HOSTNAME DISKL ORIG-KB OUT-KB COMP% MMM:SS KB/s MMM:SS KB/s -- - cable/lib1 37290 14112 37.8 0:37 376.9 0:26 538.1 cable/root 0 320288 90.0 0:01 413.7 0:07 48.0 cable-lib/amanda 0 40 32 80.0 0:00 181.0 0:02 34.9 leto /home/toto 0 20 32 160.0 0:01 51.0 0:03 21.0 with level(amoverview): host disk ...05 24 cable/lib 0 cable/root 0 cable/var/lib/amanda0 leto /home/toto 0 I understand that amanda needs more tape to realize incremental backup with this configuration, but isn't it possible to make a tape with a full backup and incremental of a same disk in the same tape??? Actually i suppose when amanda reuse a tape, it simply erase it, amanda does this if tape contains other disks for an another configuration (ex: DailySet2)??? Beside, i reaffirm a little bug with amrecover: >EOF, check amidxtaped.debug file on localhost. >amrecover: short block 0 bytes >UNKNOWN file >amrecover: Can't read file header >extract_list - child returned non-zero status: 1 and little : 'mt rewind' is adviced before OR retry a 2nd extract with amrecover -> that's works well. I didn't found into amanda commands for deleting a disk into a tapes(not into the database definitely), is it possible with amanda or other command??? Last point, i didn't found in the archive the user of hardware compression in the DI-30, anybody can manage this compression???(i just find an untested command: "mt compression) thank's by advance for yours responses. PS: sorry for my english level __ ifrance.com, l'email gratuit le plus complet de l'Internet ! vos emails depuis un navigateur, en POP3, sur Minitel, sur le WAP... http://www.ifrance.com/_reloc/email.emailif
Re: reuse tape issue
Thanks for the help. I had my tapecycle set too high. I want to say that this is one of the most helpful lists I have ever been subscribed to. Thanks again. Andrew On Wed, 16 May 2001, Chris Marble wrote: > [EMAIL PROTECTED] wrote: > > > > This morning I needed to overwrite an old tape. I looked in the tapelist > > file and my tape is listed as reuse. If I try to flush data to that tape > > I get the error "cannot overwrite active tape". Must I run an amrmtape > > before I can reuse a tape? I feel I am missing some piece of this puzzle, > > so if one could shed some light. Thanks again in advance for your help. > > In you amanda.conf file you specify tapecycle. Using that and dumpcycle > and runspercycle Amanda figures out how often you should reuse a tape. > If you lower tapecycle then you should be able to reuse this tape without > problems. > -- > [EMAIL PROTECTED] - HMC UNIX Systems Manager > My opinions are my own and probably don't represent anything anyway.. >
Lev 0 backup to disk - single system
Does anyone know how to get amanda to run a full backup but leave the dumpfiles on the holding disk after writing to tape? Holding disk space is not a problem. I want to do this for running a full backup on a single system prior to hardware upgrades, but save the time spent pulling dumps back off from tape when restoring afterward. Just trying to streamline a process to make life easier... Thanks... David Carter McLeodUSA Information Systems [EMAIL PROTECTED] 281-465-1835
client timing out
Title: client timing out I installed Amanda and I'm trying to backup the client (also the Amanda host). I am getting the message that the client is timing out. Any suggesstions? Thanks Anthony Carter
RE: reuse tape issue
'reuse' just means that the tape is available. AMANDA will want to use the oldest 'reuse' tape in the backup set for the next backup. IIRC, that's the last tape in tapelist. If you run the amcheck command, it will tell you what tape AMANDA is expecting next. If you use that tape, you will not need to use amrmtape, which is generally used for renaming tapes or removing tapes which are no longer usable for backup or restore (streched, melted, crushed, etc.) -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]] Sent: Wednesday, May 16, 2001 10:32 AM To: [EMAIL PROTECTED] Subject: reuse tape issue Hello, This morning I needed to overwrite an old tape. I looked in the tapelist file and my tape is listed as reuse. If I try to flush data to that tape I get the error "cannot overwrite active tape". Must I run an amrmtape before I can reuse a tape? I feel I am missing some piece of this puzzle, so if one could shed some light. Thanks again in advance for your help. Andrew Hall
Re: reuse tape issue
[EMAIL PROTECTED] wrote: > > This morning I needed to overwrite an old tape. I looked in the tapelist > file and my tape is listed as reuse. If I try to flush data to that tape > I get the error "cannot overwrite active tape". Must I run an amrmtape > before I can reuse a tape? I feel I am missing some piece of this puzzle, > so if one could shed some light. Thanks again in advance for your help. In you amanda.conf file you specify tapecycle. Using that and dumpcycle and runspercycle Amanda figures out how often you should reuse a tape. If you lower tapecycle then you should be able to reuse this tape without problems. -- [EMAIL PROTECTED] - HMC UNIX Systems Manager My opinions are my own and probably don't represent anything anyway.
Device ..
Title: Clear Day Folks, How would I direct Amanda to make a backup on a disk on the system? Dr Prabhakar Ganapathy ( Dr.Prab)Unix Systems AdminSNET Internet Services203-985-4724
reuse tape issue
Hello, This morning I needed to overwrite an old tape. I looked in the tapelist file and my tape is listed as reuse. If I try to flush data to that tape I get the error "cannot overwrite active tape". Must I run an amrmtape before I can reuse a tape? I feel I am missing some piece of this puzzle, so if one could shed some light. Thanks again in advance for your help. Andrew Hall
