Re: kernels v2.6.x vs. /dev/nst0 ???
Hi, Michael, on Montag, 08. März 2004 at 06:10 you wrote to amanda-users: MDS Amanda 2.4.4p1 MDS Debian Linux MDS I finally took the plunge yesterday, and upgraded my last system to MDS kernel 2.6.3. MDS What do you think? Just for the records: I compiled and ran various versions and snapshots of AMANDA on all Linux-kernels since 2.5.79 on my testbox. No problems. Right now I have 2.6.4-rc2 up, with AMANDA 2.4.5b1 on it, running fine. As Frank said, it is likely that you miss a module. How did you configure 2.6.3? A good start would be to take the config of your 2.4.25, make oldconfig and start testing with this. After that you can still try to add/remove features from you kernel. Can you access your 2.6.3-tape-device via mt or create archives on it with tar? -- best regards, Stefan Stefan G. Weichinger mailto:[EMAIL PROTECTED]
Re: Strange DNS lookup problems ... I think ...
Hi, Geoff, on Montag, 08. März 2004 at 00:26 you wrote to amanda-users: GS Here's the output from the weekend: GS Subject:schedule6 AMANDA MAIL REPORT FOR March 5, 2004 GSDate: Fri, 5 Mar 2004 21:11:03 +1100 (EST) GSFrom:Amanda Archiving Server [EMAIL PROTECTED] GS To: [EMAIL PROTECTED] In your first email you wrote: The amanda host machine is windamere Why does your report come from dipnr.nsw.gov.au then? GS These dumps were to tape schedule6-WEEK2. GS The next tape Amanda expects to use is: schedule6-WEEK3. GS FAILURE AND STRANGE DUMP SUMMARY: GS intrap /htdocs lev 0 FAILED [ [host GS windamere.dlwc.nsw.gov.au: hostname GS lookup failed]] Please show us your disklist entry for this one. GS NOTES: GS planner: tapecycle (6) = runspercycle (6) You added two tapes, but you still have these two parameters on the same value. tapecycle should be BIGGER THAN runspercycle. GS planner: Last full dump of intrap:/htdocs on tape overwritten in 1 run. Lost your last full dump here ... GS intrap /appl 0 12363201236320 --3:395636.7 6:493020.1 GS intrap /htdocs 0 FAILED GS --- As one DLE works and another on the same host does not work, it is very likely that the setup of the DLE is not right. Show us your disklist and the relevant dumptypes. Did this one ever work? -- best regards, Stefan Stefan G. Weichinger mailto:[EMAIL PROTECTED]
Re: kernels v2.6.x vs. /dev/nst0 ???
On Sun, 7 Mar 2004 at 11:10pm, Michael D Schleif wrote Amanda 2.4.4p1 Debian Linux I finally took the plunge yesterday, and upgraded my last system to kernel 2.6.3. Unfortunately, this is my amanda backup server, and under this kernel it no longer communicates with /dev/nst0: # sudo mt -f /dev/nst0 offline mt: No such device. Cannot open '/dev/nst0'. In addition to making sure st is available, doesn't 2.6 also have a device mapper (like devfs, but not) available? Did you compile that in, which would have the result of putting /dev/nst0 somewhere else (like /dev/tape/0 or something like that)? Just guessing -- I haven't played with 2.6 yet. -- Joshua Baker-LePain Department of Biomedical Engineering Duke University
Re: CHS tape robot management software
On Mon, 8 Mar 2004 at 5:06pm, Geoff Swavley wrote Does anyone have the CHS source (I would not ask normally but my vendor is unable to supply it). This is what I have: It seems to me that now would be a good time to look into alternative methods of controlling the robot. Check out mtx (and thus chg-zd-mtx), and/or look into chg-scsi. -- Joshua Baker-LePain Department of Biomedical Engineering Duke University
Re: kernels v2.6.x vs. /dev/nst0 ???
* Joshua Baker-LePain [EMAIL PROTECTED] [2004:03:08:08:18:07-0500] scribed: On Sun, 7 Mar 2004 at 11:10pm, Michael D Schleif wrote Amanda 2.4.4p1 Debian Linux I finally took the plunge yesterday, and upgraded my last system to kernel 2.6.3. Unfortunately, this is my amanda backup server, and under this kernel it no longer communicates with /dev/nst0: # sudo mt -f /dev/nst0 offline mt: No such device. Cannot open '/dev/nst0'. In addition to making sure st is available, doesn't 2.6 also have a device mapper (like devfs, but not) available? Did you compile that in, which would have the result of putting /dev/nst0 somewhere else (like /dev/tape/0 or something like that)? Yes, that's what I'm guessing. I had to add two (2) modules to get my mouse working under X11. So far, I haven't found reference to anything similar for st . . . -- Best Regards, mds mds resource 877.596.8237 - Dare to fix things before they break . . . - Our capacity for understanding is inversely proportional to how much we think we know. The more I know, the more I know I don't know . . . -- signature.asc Description: Digital signature
Re: kernels v2.6.x vs. /dev/nst0 ???
* Michael D Schleif [EMAIL PROTECTED] [2004:03:07:23:10:40-0600] scribed: snip / Unfortunately, this is my amanda backup server, and under this kernel it no longer communicates with /dev/nst0: # sudo mt -f /dev/nst0 offline mt: No such device. Cannot open '/dev/nst0'. Everything else appears to work as expected. I didn't notice this tape problem until this morning, I saw the backup was stuck on holdingdisk. Running following results in `0' tape devices found: # sudo /etc/init.d/mt-st modload Regardless of kernel, I see this: # cat /proc/scsi/scsi Attached devices: Host: scsi0 Channel: 00 Id: 05 Lun: 00 Vendor: HP Model: C1537A Rev: L812 Type: Sequential-AccessANSI SCSI revision: 02 . . . Now, after reboot back into 2.4.25, /dev/nst0 is working fine, and I am flushing last night's backup to tape as I compose this. snip / * Stefan G. Weichinger [EMAIL PROTECTED] [2004:03:08:10:12:46+0100] scribed: Hi, Michael, on Montag, 08. März 2004 at 06:10 you wrote to amanda-users: MDS Amanda 2.4.4p1 MDS Debian Linux MDS I finally took the plunge yesterday, and upgraded my last system to MDS kernel 2.6.3. MDS What do you think? Just for the records: I compiled and ran various versions and snapshots of AMANDA on all Linux-kernels since 2.5.79 on my testbox. No problems. Right now I have 2.6.4-rc2 up, with AMANDA 2.4.5b1 on it, running fine. As Frank said, it is likely that you miss a module. How did you configure 2.6.3? A good start would be to take the config of your 2.4.25, make oldconfig and start testing with this. I am using the stock Debian kernel: kernel-image-2.6.3-1-686 From config-2.6.3-1-686: # # SCSI device support # CONFIG_SCSI=m CONFIG_SCSI_PROC_FS=y # # SCSI support type (disk, tape, CD-ROM) # CONFIG_BLK_DEV_SD=m CONFIG_CHR_DEV_ST=m CONFIG_CHR_DEV_OSST=m CONFIG_BLK_DEV_SR=m # CONFIG_BLK_DEV_SR_VENDOR is not set CONFIG_CHR_DEV_SG=m # # Some SCSI devices (e.g. CD jukebox) support multiple LUNs # CONFIG_SCSI_MULTI_LUN=y # CONFIG_SCSI_REPORT_LUNS is not set CONFIG_SCSI_CONSTANTS=y CONFIG_SCSI_LOGGING=y From config-2.4.25-1-686: # # SCSI support type (disk, tape, CD-ROM) # CONFIG_BLK_DEV_SD=m CONFIG_SD_EXTRA_DEVS=40 CONFIG_CHR_DEV_ST=m CONFIG_CHR_DEV_OSST=m CONFIG_BLK_DEV_SR=m # CONFIG_BLK_DEV_SR_VENDOR is not set CONFIG_SR_EXTRA_DEVS=2 CONFIG_CHR_DEV_SCH=m CONFIG_CHR_DEV_SG=m # # Some SCSI devices (e.g. CD jukebox) support multiple LUNs # # CONFIG_SCSI_DEBUG_QUEUES is not set CONFIG_SCSI_MULTI_LUN=y CONFIG_SCSI_CONSTANTS=y CONFIG_SCSI_LOGGING=y What am I missing? After that you can still try to add/remove features from you kernel. Can you access your 2.6.3-tape-device via mt or create archives on it with tar? That's the point, I cannot (see above). -- Best Regards, mds mds resource 877.596.8237 - Dare to fix things before they break . . . - Our capacity for understanding is inversely proportional to how much we think we know. The more I know, the more I know I don't know . . . -- signature.asc Description: Digital signature
Re: kernels v2.6.x vs. /dev/nst0 ???
Hi, Michael, on Montag, 08. März 2004 at 15:39 you wrote to amanda-users: MDS * Joshua Baker-LePain [EMAIL PROTECTED] MDS [2004:03:08:08:18:07-0500] scribed: In addition to making sure st is available, doesn't 2.6 also have a device mapper (like devfs, but not) available? Did you compile that in, which would have the result of putting /dev/nst0 somewhere else (like /dev/tape/0 or something like that)? MDS Yes, that's what I'm guessing. I had to add two (2) modules to get my MDS mouse working under X11. So far, I haven't found reference to anything MDS similar for st . . . As I said before, try to make your 2.4-config first. You also have to adjust your modutils and such, but this is off-topic here. -- best regards, Stefan Stefan G. Weichinger mailto:[EMAIL PROTECTED]
Re: Solaris 8 and L200
I'm running Solaris 8, a Powerstor L200 with Amanda version 2.4.4p2. I compiled the software on the machine. It is a scsi tape changer. Thanks Christoph Scheeder wrote: Hi, first: please keep this task on the list, it gives other people the posibility to jump in and correct me if i tell you something wrong I guess this is a scsi-tapechanger, correct? then you have a few possibilities to get the robot moving: chg-scsi, mtx, and possibly more, i don't know all the free available tools to make tape robots move, espacialy for solaris as i am using linux on all my servers and i opperate my tapechanger all with chg-scsi. so now we are at the point where one of the solaris-people here on the list should jump in to help you further, but the need a little more info on your problem likthe exact type of your changer, which version of amanda you use, if you compiled it your self etc. Christoph Glenn Zenker schrieb: Thanks, that is a big help! I'll switch to another script and see what happens. Do you know if Amanda can move the robot?? Currently, the only software we have works through a webpage. Thanks, Glenn Christoph Scheeder wrote: Hi, You have configured amanda to use chg-multi as changer script. i don't think this is what you want. chg-multi is used with multiple tapedrives to form a virtual tape-robot. use one of the other chg-scripts, depending on which software you use to move your robot. Christoph Glenn Zenker schrieb: I have amanda installed when I run the following command as the user amanda /usr/local/sbin/amverify DailySet1 0 I get this: Tapes: Errors found: amtape: could not load slot 1: chg-multi: slot is empty amverify DailySet1 Thu Mar 4 13:18:52 EST 2004 Loading 1 slot... ** Error loading slot 1 amtape: could not load slot 1: chg-multi: slot is empty If anyone can lend a hand with how they configured their L200, that would be greatly appreciated.
Re: kernels v2.6.x vs. /dev/nst0 ???
Hi, Michael, on Montag, 08. März 2004 at 15:48 you wrote to amanda-users: MDS I am using the stock Debian kernel: MDSkernel-image-2.6.3-1-686 MDS From config-2.6.3-1-686: MDS# MDS# SCSI device support MDS# MDSCONFIG_SCSI=m MDSCONFIG_SCSI_PROC_FS=y MDS# MDS# SCSI support type (disk, tape, CD-ROM) MDS# MDSCONFIG_BLK_DEV_SD=m MDSCONFIG_CHR_DEV_ST=m ... MDS What am I missing? Maybe you don't have the right modutils installed. The format of modules has changed with 2.6, so you need other tools for using and handling them. Are you able to do a modprobe scsi in your 2.6-environment? Do the commands modprobe, lsmod, rmmod work there? -- best regards, Stefan Stefan G. Weichinger mailto:[EMAIL PROTECTED]
Re: Solaris 8 and L200
I don't have an L200 I have an L9 and I'm running Solaris 9, but I
assume the setup will be pretty much the same. I use mtx-1.3.8, I don't
recall where I downloaded it from however a google search should pick it
up easely. You must make sure the sgen driver is configured in your
kernel so check /kernel/drv/sgen.conf, this should give you an entry in
/dev/scsi/changer, mine is:
c3t0d0 - ../../../devices/[EMAIL PROTECTED],4000/[EMAIL PROTECTED],1/[EMAIL PROTECTED],0:changer
A search of the archives turned up a number of scripts which purport to
work with amanda and mtx, I picked the one that seemed to have been
modified most recently: mtx-changer. I have attached my copy with my
configuration.
I amanda.conf I have the following:
tpchanger mtx-changer # the tape-changer glue script
tapedev /dev/rmt/0hn # the no-rewind tape device to be used
rawtapedev /dev/null # the raw device to be used (ftape only)
changerfile
changerdev /dev/scsi/changer/c3t0d0
the mtx-changer script itself resides in /usr/local/libexec.
hope this helps.
rgds Alastair
Christoph Scheeder wrote:
Hi,
first: please keep this task on the list, it gives other people the
posibility to jump in and correct me if i tell you something wrong
I guess this is a scsi-tapechanger, correct? then you have a few
possibilities to get the robot moving:
chg-scsi, mtx, and possibly more, i don't know all the free available
tools to make tape robots move, espacialy for solaris as i am using
linux on all my servers and i opperate my tapechanger all with chg-scsi.
so now we are at the point where one of the solaris-people here on the
list should jump in to help you further, but the need a little more info
on your problem likthe exact type of your changer, which version of
amanda you use, if you compiled it your self etc.
Christoph
Glenn Zenker schrieb:
Thanks, that is a big help! I'll switch to another script and see
what happens.
Do you know if Amanda can move the robot?? Currently, the only
software we have works through a webpage.
Thanks,
Glenn
Christoph Scheeder wrote:
Hi,
You have configured amanda to use chg-multi as changer script.
i don't think this is what you want. chg-multi is used with
multiple tapedrives to form a virtual tape-robot.
use one of the other chg-scripts, depending on which software
you use to move your robot.
Christoph
Glenn Zenker schrieb:
I have amanda installed when I run the following command as the
user amanda
/usr/local/sbin/amverify DailySet1 0
I get this:
Tapes:
Errors found:
amtape: could not load slot 1: chg-multi: slot is empty
amverify DailySet1
Thu Mar 4 13:18:52 EST 2004
Loading 1 slot...
** Error loading slot 1
amtape: could not load slot 1: chg-multi: slot is empty
If anyone can lend a hand with how they configured their L200, that
would be greatly appreciated.
#! /bin/sh
###
# AMANDA Tape Changer script for use with the MTX tape changer program
# Version 1.1 - Wed Mar 7 16:50:05 CST 2001
#
# Based on 'stc-changer' by Eric Berggren ([EMAIL PROTECTED])
# Updated by Tim Skirvin ([EMAIL PROTECTED])
#
# Given that there's no license...let's make this the Perl Artistic License.
# Just make sure you give me and Eric credit if you modify this.
###
### USER CONFIGURATION
# Name of the tape drive (takes place of tapedev option in amanda.conf)
# and default driver number in library (usu 0) that DRIVE_NAME points to
DRIVE_NAME=/dev/rmt/0hn # /dev/rmt/0n ?
DRIVE_NUM=0
# Location of STC command and control device
MTX_CMD=/usr/local/sbin/mtx;
MTX_CONTROL=/dev/scsi/changer/c3t0d0;
# Whether tape drive must eject tape before changer retrieves
# (ie, EXB-2x0). Usually okay if set while not necessary, bad if
# required but not set.
DRIVE_MUST_EJECT=1
# How long to check drive readiness (in seconds) after mounting (or
# ejecting) a volume (on some libraries, the motion or eject command may
# complete before the drive has the volume fully mounted and online,
# or ready for retrieval, resulting in Drive not ready/Media not
# ready errors). Do an mt status command every 5 seconds upto this
# time.
DRIVE_READY_TIME_MAX=60
# tape mt command location...
MT_CMD=/usr/bin/mt # called via MT_CMD -f DRIVE_NAME rewind
# MT_CMD -f DRIVE_NAME offline to eject
# and MT_CMD -f DRIVE_NAME status to get ready info
##
#
NumDrives=-1
NumSlots=-1
LastSlot=-1
LoadedTape=-1
#
# Usage information
#
usage()
{
echo
echo Usage: $Progname command [arg...]
echo -info reports capability and loaded tape
echo -slot slot loads specified tape into drive
echo current reports current mounted tape
echo next loads logically next tape (loops to
User size caps
Currently running Amanda for backups of desktop shares. The problem we come across is amount of tape used by certain individuals (mostly of the 'managerial persuasion'). We have asked these individuals to limit the size of data kept in their shares many times, but they always 'creep up' above this size, to the point that they start affecting the other backups, come their rotation for a full backup. Is there any way to put a 'size cap' on the amount of data being backed up from one particular source (or -all- sources, at that)? There's a big difference between 'Could you keep the size of the share you want backed up to X?' and 'There is a limit to back up sizes of X. Anything you have above this WILL NOT be backed up...' -- Daniel Bentley - Network Technician, QSI Corporation (www.qsicorp.com) Exploits care not whence the clicks come...
Re: User size caps
On Mon, 8 Mar 2004 at 11:29am, Daniel Bentley wrote Currently running Amanda for backups of desktop shares. The problem we come across is amount of tape used by certain individuals (mostly of the 'managerial persuasion'). We have asked these individuals to limit the size of data kept in their shares many times, but they always 'creep up' above this size, to the point that they start affecting the other backups, come their rotation for a full backup. Is there any way to put a 'size cap' on the amount of data being backed up from one particular source (or -all- sources, at that)? There's a big difference between 'Could you keep the size of the share you want backed up to X?' and 'There is a limit to back up sizes of X. Anything you have above this WILL NOT be backed up...' Sounds like a job for quotas to me. I can't really think of a way of enforcing this with amanda. -- Joshua Baker-LePain Department of Biomedical Engineering Duke University
Re: User size caps
* Joshua Baker-LePain ([EMAIL PROTECTED]) [040308 11:48]: On Mon, 8 Mar 2004 at 11:29am, Daniel Bentley wrote Is there any way to put a 'size cap' on the amount of data being backed up from one particular source (or -all- sources, at that)? There's a big difference between 'Could you keep the size of the share you want backed up to X?' and 'There is a limit to back up sizes of X. Anything you have above this WILL NOT be backed up...' Sounds like a job for quotas to me. I can't really think of a way of enforcing this with amanda. The only thing I could suggest would be dynamically generating an excludes file on the target based on disk space utilization. Not exactly built-in to amanda, but a little less harsh than quotas (although I wouldn't want to be the system administrator (or his boss) when one of the managerial types loses a file and can't get it back...). Best, -- Patrick Michael Kane [EMAIL PROTECTED]
Re: User size caps
Isn't this situation better suited by implementing disk quota ? Unfortunately, the sources in question are Windows shares based on desktop machines. Hence our real, 'direct' control is limited to the server side (via. some kind of 'size throttling' through Amanda and/or SMB), or in creating a special 'shared backup partition' on each machine, size regulation coming from the physical size of the partition. The idea of 'size throttling' would be much more dynamic than physical partitions, but it's the 'kludgibility' of such that doesn't look too promising thusfar... -- Daniel Bentley - Network Technician, QSI Corporation (www.qsicorp.com) Exploits care not whence the clicks come...
[Solved]Re: Resend: ? path of config file for amrecover
I finally figure our the cause of the problem is that in /etc/inetd.conf, the entries there still pointed to old location. So each time amrecover started, inetd triggered the old daemon so that it tried to find something in old directories. Thanks so much. Allen Liu IP Application Design and Engineering Bell Canada (613) 781-7368, [EMAIL PROTECTED] 1240 -160 Elgin St, Ottawa,ON, K2P 2C4 - Original Message - From: Christoph Scheeder [EMAIL PROTECTED] To: Allen Liu --- work [EMAIL PROTECTED] Cc: Paul Bijnens [EMAIL PROTECTED]; [EMAIL PROTECTED] Sent: Saturday, March 06, 2004 3:57 AM Subject: Re: Resend: ? path of config file for amrecover Hi, i'll bet you have at least 2 versions of amrecover lying around on your system, and the first one it finds in the path is the old one... if you change the install-path you'll be responsible for deleting the old installation, make install can't do that for you Christoph Allen Liu --- work schrieb: - Original Message - From: Paul Bijnens [EMAIL PROTECTED] To: Allen Liu --- work [EMAIL PROTECTED] Cc: [EMAIL PROTECTED] Sent: Friday, March 05, 2004 12:06 PM Subject: Re: Resend: ? path of config file for amrecover Allen Liu --- work wrote: Just resend since I haven't heard anything back on it. I installed amanda in my Solaris box with configuration 'snd' like : /myapp/am1/etc/amanda/snd But when I run amrecover , it tried to to search config file from my old installation path /myapp/am/etc/amanda ( see below), it ceitainly failed. I wonder why it tried that way ? and how to make it search my new installation location ? If you change the install path, you'll have to recompile. Yes, I re-did everything : configure/make/make install. But it still behaves like this. I will try to delete everything and reinstall it. I just wonder why. -- Paul Bijnens, XplanationTel +32 16 397.511 Technologielaan 21 bus 2, B-3001 Leuven, BELGIUMFax +32 16 397.512 http://www.xplanation.com/ email: [EMAIL PROTECTED] *** * I think I've got the hang of it now: exit, ^D, ^C, ^\, ^Z, ^Q, F6, * * quit, ZZ, :q, :q!, M-Z, ^X^C, logoff, logout, close, bye, /bye, * * stop, end, F3, ~., ^]c, +++ ATH, disconnect, halt, abort, hangup, * * PF4, F20, ^X^X, :D::D, KJOB, F14-f-e, F8-e, kill -1 $$, shutdown, * * kill -9 1, Alt-F4, Ctrl-Alt-Del, AltGr-NumLock, Stop-A, ...* * ... Are you sure? ... YES ... Phew ... I'm out * ***
Re: kernels v2.6.x vs. /dev/nst0 ???
* Stefan G. Weichinger [EMAIL PROTECTED] [2004:03:08:16:03:45+0100] scribed: Hi, Michael, on Montag, 08. März 2004 at 15:48 you wrote to amanda-users: MDS I am using the stock Debian kernel: MDSkernel-image-2.6.3-1-686 MDS From config-2.6.3-1-686: MDS# MDS# SCSI device support MDS# MDSCONFIG_SCSI=m MDSCONFIG_SCSI_PROC_FS=y MDS# MDS# SCSI support type (disk, tape, CD-ROM) MDS# MDSCONFIG_BLK_DEV_SD=m MDSCONFIG_CHR_DEV_ST=m ... MDS What am I missing? Maybe you don't have the right modutils installed. The format of modules has changed with 2.6, so you need other tools for using and handling them. Are you able to do a modprobe scsi in your 2.6-environment? Do the commands modprobe, lsmod, rmmod work there? For the record, upgrading module-init-tools from 3.0-pre9-1 to 3.0-pre10-1 did the trick for me! I do not know why this worked, because the changelog indicates only some upstream release issue. Also, as a loadable module (regardless kernel 2.4.25 or 2.6.3), `st' is _not_ loaded (e.g., lsmod) until userland tries to use it. In my case, that is the `mt-st' package, which I have not changed in many weeks. Bottomline, I upgraded module-init-tools and rebooted, and now userland has no problems communicating with /dev/nst0. Thank you for your help. -- Best Regards, mds mds resource 877.596.8237 - Dare to fix things before they break . . . - Our capacity for understanding is inversely proportional to how much we think we know. The more I know, the more I know I don't know . . . -- signature.asc Description: Digital signature
Re: Strange DNS lookup problems ... I think ...
I read the last e-mail about this, but lost it, but I think I remember the basic details. First, I would try setting up some sort of nameservice caching on the client and server as a work-around. Some flavors of Linux have a caching-nameserver package that sets up the correct bind files for you, then you just put nameserver 127.0.01 at the top of /etc/resolv.conf. tmdns is supposed to be a more lightweight caching nameserver of some sort, but I haven't had good luck with it so far. nscd is a more general-purpose nameservice caching mechanism that can also cache NIS and LDAP data, but I think there may be a kernel piece to it that you also need compiled into the kernel. SGI IRIX has nsd which is similar to nscd. If you use nscd or nsd, check /etc/nsswitch.conf for the order that name services will be checked for hosts. In particular, you may need to delete nis+ or nisplus if you don't have NIS+ running on your network--It is often in there as one of the defaults, but can cause the host res process to crap out at that point if you don't have NIS+ available on your network. Second, I would check interface statistics on the client, server, nameserver, and switches and routers if possible. You want to check for collisions and/or errors, and keep an eye out for duplex mismatch or auto-negotiation problems related to certain hardware. Watch out for misbehaving mini-hubs or mini-switches along the way. I have had problems with interface hotplug on Linux and certain cards not detecting a link or auto-negotiating correctly, eg. 3c509B. I had to put MII_NOT_SUPPORTED=yes in /etc/sysconfig/network-scripts/ifcfg-ethX where X is the number of the interface, to explicitly disable hotplug for that adapter. --jonathan
hybrid theory
Is anyone successfully using a mixed strategy of backups to both disk and tape? In particular, I have a 1 TB Snap 4500 and an LTO tape drive. I have thought about a few different ways to go about it, but would appreciate suggestions from someone who has tried this approach. The Snap also has built-in snapshot capability, and I wondered if there was a way to make use of that in combination with backups to disk. 1. normal amanda backup to holding disk, flush when disk is getting full 2. do full dumps to tape out of cycle and incrementals to disk with file-driver 3. #2 but also use snapshot technology to keep an even longer history of incrementals 4. use file-driver, and occasionally archive some of the tape-files to real tape Any other ideas? Thanks, --jonathan
Re: Solaris 8 and L200
I currently use CHS with solaris (all versions 2.6 - 2.8) ... but as I have been discussing with the list I only have binaries to run it in 32bit mode. I can keep you posted Glenn on what I change to . I have had a quick look at MTX and I am still trying to find chg-scsi! Chs works well ... and if I had a preference I would continue using it as our amanda setup with CHS has worked fine for 5 years but no-one seems to have the source code for a recompile into 64bit mode so I'll start looking at these others. Let you know how I go. Glenn Zenker wrote: I'm running Solaris 8, a Powerstor L200 with Amanda version 2.4.4p2. I compiled the software on the machine. It is a scsi tape changer. Thanks Christoph Scheeder wrote: Hi, first: please keep this task on the list, it gives other people the posibility to jump in and correct me if i tell you something wrong I guess this is a scsi-tapechanger, correct? then you have a few possibilities to get the robot moving: chg-scsi, mtx, and possibly more, i don't know all the free available tools to make tape robots move, espacialy for solaris as i am using linux on all my servers and i opperate my tapechanger all with chg-scsi. so now we are at the point where one of the solaris-people here on the list should jump in to help you further, but the need a little more info on your problem likthe exact type of your changer, which version of amanda you use, if you compiled it your self etc. Christoph Glenn Zenker schrieb: Thanks, that is a big help! I'll switch to another script and see what happens. Do you know if Amanda can move the robot?? Currently, the only software we have works through a webpage. Thanks, Glenn Christoph Scheeder wrote: Hi, You have configured amanda to use chg-multi as changer script. i don't think this is what you want. chg-multi is used with multiple tapedrives to form a virtual tape-robot. use one of the other chg-scripts, depending on which software you use to move your robot. Christoph Glenn Zenker schrieb: I have amanda installed when I run the following command as the user amanda /usr/local/sbin/amverify DailySet1 0 I get this: Tapes: Errors found: amtape: could not load slot 1: chg-multi: slot is empty amverify DailySet1 Thu Mar 4 13:18:52 EST 2004 Loading 1 slot... ** Error loading slot 1 amtape: could not load slot 1: chg-multi: slot is empty If anyone can lend a hand with how they configured their L200, that would be greatly appreciated. -- geoffs :-) -- Geoff Swavley Email : [EMAIL PROTECTED] UNIX Sys Admin Snail : Level 8, 10 Valentine Ave, Support and Network Operations Parramatta NSW 2150 Dept of Infrastructure, PlanningSydney Australia and Natural Resources Phone : 61-2-98957125 http://www.radx.net/~geoffs Fax : 61-2-98957086 Mobile: 61-422-002005 Home : 61-2-96593637 -- Be wary of the man who urges an action in which he himself incurs no risk. - Setanti, Joaquin de
Re: hybrid theory
On Mon, Mar 08, 2004 at 05:35:21PM -0500, Jonathan Dill wrote: Is anyone successfully using a mixed strategy of backups to both disk and tape? In particular, I have a 1 TB Snap 4500 and an LTO tape drive. I have thought about a few different ways to go about it, but would appreciate suggestions from someone who has tried this approach. The Snap also has built-in snapshot capability, and I wondered if there was a way to make use of that in combination with backups to disk. 1. normal amanda backup to holding disk, flush when disk is getting full 2. do full dumps to tape out of cycle and incrementals to disk with file-driver 3. #2 but also use snapshot technology to keep an even longer history of incrementals 4. use file-driver, and occasionally archive some of the tape-files to real tape Any other ideas? Set up normal backups to tape, probably already done. Set up the file:driver to use the Snap server Set up the RAIT driver to mirror the backups to two devices, one the file:driver the other the LTO. Use the LTO for longterm storage, the Snap server for immediate availability of backups for recovery. -- Jon H. LaBadie [EMAIL PROTECTED] JG Computing 4455 Province Line Road(609) 252-0159 Princeton, NJ 08540-4322 (609) 683-7220 (fax)
Solaris ACLs
Someone on the sage-members list is looking for free backup software that met his listed requirements, and I was about to reply with Amanda, but I wasn't sure about his requirement #5 (below) pertaining to Solaris ACLs. Will Amanda actually do what he wants? Thanks, Frank I currently work for a large university. We currently have a Very Large tape library system (a walk-in model, terrabites of storage, hundreds of tapes, blah blah blah) We currently have robotics software (unitree) and some interesting home-grown custom jobbies. Unfortunately, they are rather ugly. So I'd like to be able to migrate our backups to something a little more sane, and a little more widely used. Given the amount of data, and number of hosts, and our limited funding for software, getting a license for veritas or legato backup software, etc. is going to be out of the question. So I need a free solution suggested. My ideal backup solution would handle: 1. multiple incoming backup streams, ideally multiplexing then to a single tape or virtual restoral device, for streaming speed purposes, etc. 2a. know about interfacing with unitree directly, OR 2b. be flexible about save all the data to a pseudo-'file' which is actually managed by HSM 3. be able to handle restore requests along the lines of, Give me all the files in directory X, on machine Y, at date YYY/MM/DD:HH/MM/SS and pull in the appropriate files from the last full dump, and all relevant incrementals. And if there was a level0, level2, level3, and level4 dump, and the most recent versions of the file(s) were on level3, it would not have to go through the level0 and level2 dumps to find out the data is not there. 4. it must be able to handle Very Large Filesystems (I'm not sure we have any terrabytes filesystems... Yet. But we probably will have them soon) 5. It should be able to handle restoring Solaris ACLs It is not neccessary to have any kind of non-root-user interface. Restores are handled by the sysadmins only. Of all of the above, I think that everything except #2 is mandatory. Am I dreaming, or is there anything out there for free that can actually handle all of this? -- Frank Smith [EMAIL PROTECTED] Sr. Systems Administrator Voice: 512-374-4673 Hoover's Online Fax: 512-374-4501
Re: Solaris ACLs
On Mon, Mar 08, 2004 at 06:35:49PM -0600, Frank Smith wrote: Someone on the sage-members list is looking for free backup software that met his listed requirements, and I was about to reply with Amanda, but I wasn't sure about his requirement #5 (below) pertaining to Solaris ACLs. Will Amanda actually do what he wants? Thanks, Frank I currently work for a large university. We currently have a Very Large tape library system (a walk-in model, terrabites of storage, hundreds of tapes, blah blah blah) We currently have robotics software (unitree) and some interesting home-grown custom jobbies. Unfortunately, they are rather ugly. So I'd like to be able to migrate our backups to something a little more sane, and a little more widely used. Given the amount of data, and number of hosts, and our limited funding for software, getting a license for veritas or legato backup software, etc. is going to be out of the question. So I need a free solution suggested. My ideal backup solution would handle: 1. multiple incoming backup streams, ideally multiplexing then to a single tape or virtual restoral device, for streaming speed purposes, etc. 2a. know about interfacing with unitree directly, OR 2b. be flexible about save all the data to a pseudo-'file' which is actually managed by HSM 3. be able to handle restore requests along the lines of, Give me all the files in directory X, on machine Y, at date YYY/MM/DD:HH/MM/SS and pull in the appropriate files from the last full dump, and all relevant incrementals. And if there was a level0, level2, level3, and level4 dump, and the most recent versions of the file(s) were on level3, it would not have to go through the level0 and level2 dumps to find out the data is not there. 4. it must be able to handle Very Large Filesystems (I'm not sure we have any terrabytes filesystems... Yet. But we probably will have them soon) 5. It should be able to handle restoring Solaris ACLs It is not neccessary to have any kind of non-root-user interface. Restores are handled by the sysadmins only. Of all of the above, I think that everything except #2 is mandatory. Am I dreaming, or is there anything out there for free that can actually handle all of this? The man page doesn't mention ACL's, but I suspect it will have to preserve them. Tar/gnutar of course will not. However, if Shilly's 'star' can be made to work, it claims to preserve Solaris ACL's (and not affect atime). If ufsdump is used the normal caveats apply, exclude/include don't work, only entire file systems which must fit on tape, ... As Solaris also can do FS snapshots, the OP should be informed of that feature. Not amanda specific, but neat. -- Jon H. LaBadie [EMAIL PROTECTED] JG Computing 4455 Province Line Road(609) 252-0159 Princeton, NJ 08540-4322 (609) 683-7220 (fax)
