Re: Problem with a Seagate tape drive
"Trevor Linton" <[EMAIL PROTECTED]> writes: > I have a SEAGATE DAT (DDS-3) tape backup drive, it holds around 1 gig > of info, i'm having some errors with the amanda amdump First off, > which I think is the largest error is in the > /var/log/amanda/DailySet1/log.20010220.0, in which it gives this error > message: > > ERROR taper no-tape [rewinding tape: Input/output error] FAIL driver > fserver /home/accounting 0 [can't switch to incremental dump] > > I'm not sure what this means, the scsi device is terminated correctly, > it's on id 4 with one other device on this (hard drive) on id 1, and > id 4 does have a terminator on it. The type drive is being recognized > by the SCSI driver, in /proc/scsi/scsi it's listed as Host: SCSI1 > Channel: 00 Id: 04 Lun: 00 Vendor: ARCHIVE Model: Python 04106-XXX > Rev: 7270 Type: Sequential-Access ANSI SCSI revision:02. Any Ideas? What's your SCSI adapter? I've a similar problem with a Seagate DDS-4. It does all that I want, but the first operation after the rewind following tape insertion has to be a write. The funny thing is that I get the data and the SCSI error when I simulate amcheck. So I amlabel -f every tape with its old name directly after insertion into the drive. I use a Vortex RAID controller and I recieved the newest drivers from their support. I didn't have time to compile a new kernel yet. > The second problem and i'm not sure if this is related to the first or > not but in the same log file i have this error which is odd: This Amanda issue is treated already... [...] > I'm not subscribed to this list so if you could please respond to my > message by directly emailing me I would greatly appreciate it. I'd subscribe. You get answers for the questions you'll ask some weeks later :-) HTH, Johannes Niess
Re: HP C1553A and chio/pickers
If you are in a linux environment why dont'use mtx (and amanda chg-mtx or chg-zd-mtx) instead of chio? HP1553A (DDS-3 autoloader with 6 tape magazine?) does not have a separate scsi address for the changer, so you can use /dev/nst0 (or whatever your tape drive is indexed) as your tape device and as your changer device in amanda configuration. +---+ | Alessio Brezigar | | ARPA F-VG - OSMER | | v. Carso, 3 - loc. Scodovacca | | 33052 Cervignano d.F. (UD)| | ITALY | | Tel: +39-0431-382460 | | Fax: +39-0431-382400 | | e-mail: [EMAIL PROTECTED] | +---+ -Original Message- From: John R. Jackson <[EMAIL PROTECTED]> To: Harlan Stenn <[EMAIL PROTECTED]> Cc: [EMAIL PROTECTED] <[EMAIL PROTECTED]> Date: mercoledì 21 febbraio 2001 03:52 Subject: Re: HP C1553A and chio/pickers >>I notice: >> >> > chio params >> /dev/ch0: 6 slots, 1 drive, 0 picker >> /dev/ch0: current picker: 0 >> >>that it says it has no picker when I suspect it should say it has 1 picker. >> >>In the past I've worked around this by hacking libexec/chg-chio to ignore >>this "situation". ... >>Where is the "right" place to fix this? > >Ummm, since it's chio that's "broken", maybe that's the place to fix >the problem? > >Or are you asking about how to make an "official workaround" :-) >to Amanda? If so, just make your changes as clean as possible (i.e. >so they don't break for someone else that does not have this problem) >and post a context or unified diff, along with a ChangeLog entry and >what version of Amanda you worked against. > >>H > >John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED] >
Re: tweaking the schedule
* John R. Jackson <[EMAIL PROTECTED]> (Tue, Feb 20, 2001 at 07:53:20PM -0500) > Next we need to examine the driver algorithm in more detail. Some of > the base information is in the usenix paper on the ftp.amanda.org site. [8<] > If inparallel is set to three (or less), you don't have any "big" > dumpers and all the processing will happen from the front of the list. > If inparallel is set to four, three dumpers will be picking off the > little entries at the head of the list and one dumper will start at > the end of the list and work its way backward. Hey, that explains exactly what Im seeing ... Now that I've upped maxdumpers to 4 it is indeed picking the biggest dump first (as well as some smaller ones) and giving a reduced dumptime . Time to get that useni paper ;) Gerhard, <@jasongeo.com> == The Acoustic Motorbiker == -- __O Aristotle, Aristotle was a bugger for the bottle, =`\<, and Hobbes was fond of his Dram. (=)/(=) And Rene Descartes was a drunken fart: "I drink, therefore I am."
coherency between amindexd and amrecover
Hello, i had already a running amanda-system in former times, but i had to set up a new one. The first time, i was able to login via telnet on port 10082, so that's why i think it must be now also possible?!? I can't interpret the following lines of /tmp/amanda/amindexd.debug: amindexd: debug 1 pid 705 ruid 37 euid 37 start time Wed Feb 21 09:30:07 2001 amindexd: version 2.4.1p1 Why is it not possible to run amdump DailySet1 without running in between amcleanup? The first time, it was not required! It also seems to me that no dump is finished correctly. The amanda.conf is in the /etc/amanda/DailySet1 directory, so is that the right place because the server doesn't find it? The rights for it are set to root. Best regards, Jörn Hildebrandt -Ursprüngliche Nachricht- Von: John R. Jackson [mailto:[EMAIL PROTECTED]] Gesendet: Dienstag, 20. Februar 2001 00:11 An: Jörn Hildebrandt Cc: Amanda-Users Betreff: Re: coherency between amindexd and amrecover >saturn:~ # amrecover -C DailySet1 -s saturn -d /dev/nst0 >AMRECOVER Version 2.4.1p1. Contacting server on saturn ... >amrecover: Unexpected server end of file > >What does "Unexpectes server end.." mean?? Who produces this message?? The message comes from amrecover and means the amindexd server quit. Go to "saturn" and look in /tmp/amanda/amindexd*debug for more info. >It is also not possible to connect via telnet to the port: Why would you think it would be possible? >Jörn, John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
Re: Amrecover Problem in 2.4.2pl1
"John R. Jackson" a écrit : > > >i still have a prob with amrecover in 2.4.2pl1 > >i never have this prob in previous version > > > >i run amanda server on a Sun sol 2.6 > >... > >but when i try amrecover on the server , to restore a file or dir > >i always have : > >... > >error reading tape: Connection reset by peer > > Is there anything of interest in /tmp/amanda/amidxtaped*debug on the > tape machine? What about in wherever inetd.conf logs things? Nothing bad > > You might check out this thread, especially if your machines have > multiple network interfaces: Ok i have two network interfaces but with previous version too > > http://groups.yahoo.com/group/amanda-users/message/27372 i am not able to get the patch in the forum can you send it agoin please > > >patrick > > John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED] thanks patrick
Re: tweaking the schedule
On Tue, Feb 20, 2001 at 07:53:20PM -0500, John R. Jackson wrote: > There are two categories of dumpers, "little" and "big". The first three > (LITTLE_DUMPERS in driver.c) dumpers from "inparallel" are assigned > to "little" duty, everyone else is "big". Little dumpers pick entries > from the head (shortest estimated time) of the queue. Big dumpers take > entries from the end (longest estimated time) of the queue. I'm working on a patch where each dumpers can have different criteria for the disk it pick according to (smallest or biggest: bandwidth, size or time) The criteria I'm using might not be good for you, but it give me a good speedup. You update the type_dumpers array to change the criteria for a dumper. Jean-Louis -- Jean-Louis Martineau email: [EMAIL PROTECTED] Departement IRO, Universite de Montreal C.P. 6128, Succ. CENTRE-VILLETel: (514) 343-6111 ext. 3529 Montreal, Canada, H3C 3J7Fax: (514) 343-5834
Re: Amtape not reqinding after a label lookup
That seems to have fixed it. I tested both the 'slot' and the 'label' params, and nothing griped. Thanks... On Tue, 20 Feb 2001, John R. Jackson wrote: > >I checked the amidxtaped.debug, and I've figured out what the problem is. > >After amtape is called with either the 'label ' or 'current' or any > >other where it returns the amlabel of the tape, amtape does not rewind the > >tape when it's finished reading the amlabel. > >... Is there a way to have > >amtape automatically do the rewind after an amlabel check? ... > > That's probably a reasonable idea. Give the following (2.4.2p1) patch > a try and let me know if it works. It does a rewind for the "slot" > and "label" commands in amtape, but none of the others. The patch is > untested (other than getting a clean compile). > > > Jason Hollinden > > John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED] > > (Run "e mhn" at "What now?", then $}base64 -d as needed and remove the > Content-Transfer-Encoding line (leave it blank) and these lines.) -- Jason Hollinden SMG Systems Admin
Re: questions on DLT7?
Thinking about purchasing one want to make sure it will work before spending the large amount of money. Denyce John R. Jackson wrote: >> Does anyone know if there is a problem using the DLT7 with amanda or >> solaris 2.6 or if there are drivers for solaris 8? > > > I don't know what a "DLT7" is, but on the assumption it's a DLT7000, > they work fine at 2.6. I've got almost a dozen. > > Why? Are you having trouble? > >> Denyce > > > John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
gnutar socket ignored (although I'm using 2.4.2p1)
Hello, I'm receiving messages like the following in my daily
report:
/-- pds/var lev 1 STRANGE
sendbackup: start [pds:/var level 1]
sendbackup: info BACKUP=/usr/local/bin/gtar
sendbackup: info RECOVER_CMD=/usr/local/bin/gtar -f... -
sendbackup: info end
? gtar: ./account/acct: file changed as we read it
| gtar: ./run/log: socket ignored
| gtar: ./run/ndc: socket ignored
| Total bytes written: 12646400 (12MB, 2.1MB/s)
sendbackup: size 12350
sendbackup: end
\
I saw Alexander's post in a previous mail abut the lines:
/* GNU tar 1.13.17 will print this warning when (not) backing up a
Unix named socket. */
{ DMP_NORMAL, ": socket ignored$", 1 },
in client-src/sendbackup-gnutar.c. (I have those in there, since
that's where I copied them from.
Although I am using a FreeBSD system, I have compiled and installed
GNU tar 1.13.19, and ran ./configure with --with-gnutar=/usr/local/bin/gtar.
What else should I look at to see what is causing this to still happen?
Thanks,
-Paul
--
Paul D. Schmidt | <[EMAIL PROTECTED]>
Coder / Sys Admin |<[EMAIL PROTECTED]>
Unbound Communications | http://www.unboundcom.com
"The grass may actually be greener on the other side of the fence,
but it still has to be mowed." --Anonymous
Re: gnutar socket ignored (although I'm using 2.4.2p1)
"Paul D. Schmidt" <[EMAIL PROTECTED]> writes: > Hello, I'm receiving messages like the following in my daily > report: > > /-- pds/var lev 1 STRANGE > sendbackup: start [pds:/var level 1] > sendbackup: info BACKUP=/usr/local/bin/gtar > sendbackup: info RECOVER_CMD=/usr/local/bin/gtar -f... - > sendbackup: info end > ? gtar: ./account/acct: file changed as we read it > | gtar: ./run/log: socket ignored > | gtar: ./run/ndc: socket ignored > | Total bytes written: 12646400 (12MB, 2.1MB/s) > sendbackup: size 12350 > sendbackup: end > \ > > What else should I look at to see what is causing this to still happen? > > Thanks, > -Paul The reason you see those errors is because a file changed while it was being read. As long as there is just one line Amanda doesn't understand, you will get all messages from the backup. Lines prefixed with | are lines Amanda is ok with and those prefixed with ? are the lines it doesn't recognize. To avoid it, either make sure there is no activity on the filesystems being dumped, patch Amanda to ignore those messages (eew) or in the case it is the same files that change during every dump and they aren't important, exclude them. I do the latter with some temporary files. Hope that helps... Jens Bech Madsen -- Jens Bech Madsen The Stibo Group, Denmark
Re: questions on DLT7?
* Denyce Berg <[EMAIL PROTECTED]> (Wed, Feb 21, 2001 at 08:18:46AM -0700) > Thinking about purchasing one want to make sure it will work before > spending the large amount of money. Buy an LTO .. 100G uncompressed for roughly the same price Gerhard, (@jasongeo.com) == The Acoustic Motorbiker == -- __0 Oh where are you now pussy willow that smiled on this leaf? =`\<, When I was alone you promised the stone from your heart (=)/(=) my head kissed the ground I was half the way down, treading the sand please, please, lift a hand
Re: amreport Problem
>Hmafter typing this i see MAILER="/usr/sbin/sendmail" >OK, can i change this without new compiling amanda? You can't. That value is determined when you run ./configure. You need to "make distclean" and start over. And check the output of the ./configure to make sure it found a suitable mail program before rebuilding. Note that you only have to rebuild the server. The clients don't care about MAILER. >Juergen John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
Re: amrestore problem
>OK, i tested it with this commands: > >amanda@fileserver:~ > cd /usr/local/sbin/ >amanda@fileserver:/usr/local/sbin > ./amrestore /dev/nst0 king >amrestore: missing file header block >... >HmWhat´s wrong? Did you "mt -f /dev/nst0 rewind" first? >Juergen John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
Re: RedHat Amandad xinetd.conf etc etc
Hi Holm
I've just started with amanda on Red Hat. Spent a long night getting it working (with
a lot of help from this list).
Here's my entry in xinetd.conf... (not touched)
==
defaults
{
instances = 60
log_type= SYSLOG authpriv
log_on_success = HOST PID
log_on_failure = HOST RECORD
}
includedir /etc/xinetd.d
===
Here's the file "amanda" in the /etc/xinetd.d directory... (amanda is mentioned in
services)
===
service amanda
{
socket_type = dgram
protocol = udp
wait = yes
user = amanda
server = /usr/local/libexec/amandad
disable = no
}
The real life saver was when I added...
program"GNUTAR"
into the define dumptype in amanda.conf
Hope this helps
rgds
Dirk
- Original Message -
From: "Holm-Hansen" <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Wednesday, February 21, 2001 10:34 PM
Subject: RedHat Amandad xinetd.conf etc etc
> I'm on my last brain cell...
>
> I'm getting the dreaded "selfcheck request timed-out. Host down?" message from
> amcheck. I've read and re-read the faq. After doing an "ls -lu" in the
> server directory, I found that the executable isn't getting called when I run
> amcheck. (I believe that is what ls -lu does).
>
> in my /etc/xinetd.d directory I have the amanda file (and an identical amandad
> just in case) that looks like this:
>
> service amandad
> {
> socket_type = dgram
> protocol = udp
> wait = yes
> user = amanda
> group = amanda
> server = /usr/local/libexec/amandad
> instances = 10
> server_args = amandad
> }
>
> I don't know why I have to have "wait = yes" but I saw it in another conf
> file.
> Also I built this from source with the user=amanda and group=amanda.
>
> I realize that there are billions of that could be broken so if I need to
> elaborate more, I'd be more than happy to :)...
>
> Thanks,
> Andrew
>
>
Re: slow backups
Ben Kochie wrote:
>
> I've been wrestling with some fairly slow backups for a while now. I
> have amanda 2.4.2 server, running a dual 733 dell rack server, (256mb
> ram, UW scsi disk) with a 3com 100mb switch stack. the network is far
> from saturated.
> the only thing that gets me are the clients.. I have 3 SGI's, a
> challange-S, an O-200, and an O2. I also have a PII-450 netBSD
> machine. all run extremely slow.
> as far as I can tell from my MRTG graphs.. no single backup process will
> exceed 100KB/sec, most run around 50KB/sec
>
> here are some of my config lines...
>
> amanda.conf
> inparallel 12
> maxdumps 5
> netusage 100 Mbps
>
> define dumptype ima-ss-comp-high-tar {
> high-tar
> compress server fast
> exclude list "/usr/local/amanda/etc/amanda/Test/excludes"
> }
>
> define interface eth0 {
> comment "100 Mbps ethernet"
> use 100 Mbps
> }
>
> define interface local {
> comment "a local disk"
> use 50 mbps
> }
>
> -- disklist ---
> fs0 /home ima-ss-comp-user-tar1
> eth0
> fs0 /varsityima-ss-comp-user-tar2
> eth0
How about trying the backups without specifying the interface
parameter in the disklist? I never could get that to work the way
I wanted. I just set dumptypes with different maxdumps numbers for
different systems. My antique boxes with 10Mb EtherNet use
maxdumps 1. My fast systems have maxdumps 2.
Are you doing all the compression on the tape host? That could
grind things to a halt.
--
[EMAIL PROTECTED] - HMC UNIX Systems Manager
My opinions are my own and probably don't represent anything anyway.
Re: amrestore problem
> >amanda@fileserver:~ > cd /usr/local/sbin/ > >amanda@fileserver:/usr/local/sbin > ./amrestore /dev/nst0 king > >amrestore: missing file header block > Did you "mt -f /dev/nst0 rewind" first? i have forgotten :-( But, after rewinding the tape i become another error: root@fileserver:/usr/local/sbin > ./amrestore /dev/nst0 king amrestore: WARNING: not at start of tape, file numbers will be offset root@fileserver:/usr/local/sbin > I think i must change the tape because amanda uses two tapes per dump and at this moment the second tape is in the slot. Is this correct? BYe Juergen
Re: Questions and Answers and Thanks
Gerhard den Hollander wrote: > > Since it's a linux box , you may want to look into using reiserFS. > I found that file access with Reiser is faster than ext2 (although nothing > shocking) esp. when going through a lot of smaller files in one swoop. I formatted up the holding disk with a large blocksize and small inode count. It's never going to have over 100 files on it so I doubt ReiserFS is worth the hassle. I do all my backups with dump, ufsdump or xfsdump now. I'd just as soon avoid adding in gnu tar too. Though since I've got one Digital Unix box where the vfsdump doesn't work... -- [EMAIL PROTECTED] - HMC UNIX Systems Manager My opinions are my own and probably don't represent anything anyway.
RedHat Amandad xinetd.conf etc etc
I'm on my last brain cell...
I'm getting the dreaded "selfcheck request timed-out. Host down?" message from
amcheck. I've read and re-read the faq. After doing an "ls -lu" in the
server directory, I found that the executable isn't getting called when I run
amcheck. (I believe that is what ls -lu does).
in my /etc/xinetd.d directory I have the amanda file (and an identical amandad
just in case) that looks like this:
service amandad
{
socket_type = dgram
protocol = udp
wait = yes
user = amanda
group = amanda
server = /usr/local/libexec/amandad
instances = 10
server_args = amandad
}
I don't know why I have to have "wait = yes" but I saw it in another conf
file.
Also I built this from source with the user=amanda and group=amanda.
I realize that there are billions of that could be broken so if I need to
elaborate more, I'd be more than happy to :)...
Thanks,
Andrew
RE: RedHat Amandad xinetd.conf etc etc
Unfortunately this is nearly identical to what I have (save the disable=no
which didn't make things go). I've found that I can run amandad by hand and
it acts as predicted (timing out after 30 seconds) but it still isn't being
called from xinetd. Another curious problem is that in the faq it says to do
a
netstat -a | grep -i amanda
and that this should return something. I get nothing when I do this.
/etc/services has a listing for amanda and kamanda etc...
I'm even more lost now than I was 20 minutes ago... :)
Thanks for taking a swing at it Dirk,
If anyone else has any ideas, I'd be thrilled to hear them!
Andrew
>= Original Message From "Dirk Webster" <[EMAIL PROTECTED]>
=
>Hi Holm
>
>I've just started with amanda on Red Hat. Spent a long night getting it
working (with a lot of help from this list).
>
>Here's my entry in xinetd.conf... (not touched)
>==
>defaults
>{
>instances = 60
>log_type= SYSLOG authpriv
>log_on_success = HOST PID
>log_on_failure = HOST RECORD
>}
>
>includedir /etc/xinetd.d
>===
>Here's the file "amanda" in the /etc/xinetd.d directory... (amanda is
mentioned in services)
>===
>service amanda
>{
>socket_type = dgram
>protocol = udp
>wait = yes
>user = amanda
>server = /usr/local/libexec/amandad
>disable = no
>}
>
>The real life saver was when I added...
>
>program"GNUTAR"
>
>into the define dumptype in amanda.conf
>
>Hope this helps
>rgds
>Dirk
>
>- Original Message -
>From: "Holm-Hansen" <[EMAIL PROTECTED]>
>To: <[EMAIL PROTECTED]>
>Sent: Wednesday, February 21, 2001 10:34 PM
>Subject: RedHat Amandad xinetd.conf etc etc
>
>
>> I'm on my last brain cell...
>>
>> I'm getting the dreaded "selfcheck request timed-out. Host down?" message
from
>> amcheck. I've read and re-read the faq. After doing an "ls -lu" in the
>> server directory, I found that the executable isn't getting called when I
run
>> amcheck. (I believe that is what ls -lu does).
>>
>> in my /etc/xinetd.d directory I have the amanda file (and an identical
amandad
>> just in case) that looks like this:
>>
>> service amandad
>> {
>> socket_type = dgram
>> protocol = udp
>> wait = yes
>> user = amanda
>> group = amanda
>> server = /usr/local/libexec/amandad
>> instances = 10
>> server_args = amandad
>> }
>>
>> I don't know why I have to have "wait = yes" but I saw it in another conf
>> file.
>> Also I built this from source with the user=amanda and group=amanda.
>>
>> I realize that there are billions of that could be broken so if I need to
>> elaborate more, I'd be more than happy to :)...
>>
>> Thanks,
>> Andrew
>>
>>
