Re: Tapelist
On Tue, Mar 11, 2008 at 11:57:05AM -0500, Arthur Smith enlightened us: Jon, Thanks! I want to be sure I understand correctly: I have 10 tapes, labeled monday1, tuesday1, etc.. through friday2. I should now be able to recreate the tapelist file like so: 0 monday1 reuse 0 tuesday1 reuse and so on, and amanda will use them in the order they are put in, correct? As if they were actually new tapes (even though they're not)... Am I on track? Yes, with the exception of labeling tapes with days of the week. It's generally not recommended because amanda has no concept of what day it is and will merely ask for the next tape needed. So if you miss a day, run over onto another tape, etc., you end up with strange situations like using monday1 on a Thursday. It also makes writing the regexp in your config file harder. It is recommended to have some prefix string (e.g. Daily, Archive, Monthly) followed by incremental numbers. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: another setup problem
On Thu, Aug 16, 2007 at 02:38:03PM -0700, Brian Maxwell enlightened us: I'm sorry to be so bothersome, but I'm having another problem, and after looking through all of the documentation I could find, I can't solve it. In my disklist file, I've defined a new client, named lucy. I set up the amanda client package on lucy according to the directions in the documentation. Here is the contents of the .amandahosts file on lucy: odus amanda That is the only configuration file I set up on lucy. Should there be others? When I run amcheck back on odus (the backup server), I get this: odus 31 /home3/local/etc/amanda/normal amcheck normal Amanda Tape Server Host Check - Holding disk /home3/amanda/holding: 7297619 KB disk space available, using 7195219 KB slot 3: read label `normal104', date `X' NOTE: skipping tape-writable test Tape normal104 label ok NOTE: conf info dir /usr/local/var/amanda/normal/curinfo does not exist NOTE: it will be created on the next run. NOTE: index dir /usr/local/var/amanda/normal/index does not exist NOTE: it will be created on the next run. Server check took 0.528 seconds Amanda Backup Client Hosts Check WARNING: lucy: selfcheck request failed: timeout waiting for ACK Client check: 2 hosts checked in 30.162 seconds, 1 problem found (brought to you by Amanda 2.5.1) Odus and lucy appear in the disklist file like this: odus /dev/dsk/c0t1d0s5 hard-disk-dump # tools (/opt) lucy /dev/dsk/c0t0d0s7 hard-disk-dump # gmls (/export/home) Does anyone have any idea why amanda on odus can't talk to amanda on lucy? Thanks in advance, Did you place the correct entries in (x)inetd and restart (x)inetd? Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: amtypetype taking 51 hours??
On Sat, Jul 28, 2007 at 09:15:25PM -0500, Dustin J. Mitchell enlightened us: I have a Dell Powervault 124T LTO-2 and I'm running amtapetype on it and it is running incredibly slow. It estimates 51 hours to check the tape. I'm new to amanda, but this seems absurdly slow to me. I don't know the details of your tape device, but keep in mind that amtapetype does things that are not usual for a tape device, and that for enormous tapes, it can, indeed, take a while. Especially if you don't give it an estimate size (-e, IIRC). It should take 2xlength of time to fill one tape, which for LTO2 is around an hour, I believe. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: Troubleshooting new Amanda client: Amanda user?
On Tue, Jun 26, 2007 at 10:38:33AM -0400, Zembower, Kevin enlightened us: Kevin, thanks so much. You were right on the money. Disabling the firewall completely allow amcheck to work correctly. If you have some additional patience, I could use a hand trying to configure the firewall rules correctly on my amanda client. I tried to follow the directions at http://wiki.zmanda.com/index.php/How_To:Set_Up_iptables_for_Amanda to set up this rule on tobaccodev, my amanda client. This combines the amanda rule with the rules I set up using the firewall GUI in CentOS5 (RHEL5): [EMAIL PROTECTED] ~]# iptables -t filter -I INPUT 1 -p udp -m udp -s centernet.jhuccp.org --dport 10080:10083 -j ACCEPT [EMAIL PROTECTED] ~]# service iptables status Table: filter Chain INPUT (policy ACCEPT) num target prot opt source destination 1ACCEPT udp -- 10.253.192.205 0.0.0.0/0 udp dpts:10080:10083 2RH-Firewall-1-INPUT all -- 0.0.0.0/00.0.0.0/0 Chain FORWARD (policy ACCEPT) num target prot opt source destination 1RH-Firewall-1-INPUT all -- 0.0.0.0/00.0.0.0/0 Chain OUTPUT (policy ACCEPT) num target prot opt source destination Chain RH-Firewall-1-INPUT (2 references) num target prot opt source destination 1ACCEPT all -- 0.0.0.0/00.0.0.0/0 2ACCEPT icmp -- 0.0.0.0/00.0.0.0/0 icmp type 255 3ACCEPT esp -- 0.0.0.0/00.0.0.0/0 4ACCEPT ah -- 0.0.0.0/00.0.0.0/0 5ACCEPT udp -- 0.0.0.0/0224.0.0.251 udp dpt:5353 6ACCEPT udp -- 0.0.0.0/00.0.0.0/0 udp dpt:631 7ACCEPT tcp -- 0.0.0.0/00.0.0.0/0 tcp dpt:631 8ACCEPT all -- 0.0.0.0/00.0.0.0/0 state RELATED,ESTABLISHED 9ACCEPT tcp -- 0.0.0.0/00.0.0.0/0 state NEW tcp dpt:21 10 ACCEPT tcp -- 0.0.0.0/00.0.0.0/0 state NEW tcp dpt:25 11 ACCEPT tcp -- 0.0.0.0/00.0.0.0/0 state NEW tcp dpt:22 12 ACCEPT tcp -- 0.0.0.0/00.0.0.0/0 state NEW tcp dpt:443 13 ACCEPT tcp -- 0.0.0.0/00.0.0.0/0 state NEW tcp dpt:23 14 ACCEPT tcp -- 0.0.0.0/00.0.0.0/0 state NEW tcp dpt:80 15 REJECT all -- 0.0.0.0/00.0.0.0/0 reject-with icmp-host-prohibited Here's an example of a no-error 'amcheck -c DBackup tobaccodev' from the tapeserver: [EMAIL PROTECTED] ~]# tcpdump -nn src or dst centernet and port amanda tcpdump: verbose output suppressed, use -v or -vv for full protocol decode listening on eth0, link-type EN10MB (Ethernet), capture size 96 bytes 10:28:58.190591 IP 10.253.192.205.854 10.253.192.217.10080: UDP, length 123 10:28:58.210814 IP 10.253.192.217.10080 10.253.192.205.854: UDP, length 50 10:28:58.212936 IP 10.253.192.217.10080 10.253.192.205.854: UDP, length 87 10:28:58.214318 IP 10.253.192.205.854 10.253.192.217.10080: UDP, length 50 10:28:58.216532 IP 10.253.192.205.854 10.253.192.217.10080: UDP, length 299 10:28:58.223632 IP 10.253.192.217.10080 10.253.192.205.854: UDP, length 50 10:28:58.233581 IP 10.253.192.217.10080 10.253.192.205.854: UDP, length 527 10:28:58.235018 IP 10.253.192.205.854 10.253.192.217.10080: UDP, length 50 8 packets captured 20 packets received by filter 0 packets dropped by kernel [EMAIL PROTECTED] ~]# I had to insert the rule to allow amanda packets in _before_ the RH-Firewall-1-INPUT rule to make it work. This tests correctly with amcheck, but I haven't tried an actual dump yet. If someone with some amanda firewall rule writing experience could check and confirm my work, I'll write an addendum to the Zmanda artile with my example, for other CentOS and RHEL users. Thanks, again, Kevin, for your advice and suggestions. -Kevin On my CentOS client systems, I modify /etc/sysconfig/iptables-config to read: IPTABLES_MODULES=ip_conntrack_ftp ip_conntrack_amanda And simply allow udp 10080 from the server (in /etc/sysconfig/iptables): -A INPUT -s 192.168.1.1 -d 192.168.1.30 -p udp -m udp --dport 10080 -j ACCEPT On the server I also allow tcp 10082 and 10083. On my bridging firewall, I modify /etc/modprobe.conf to include a longer timeout: options ip_conntrack_amanda master_timeout=2400 That works for me... Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: suggestion for a disk-to-disk backup server
On Sun, Jun 24, 2007 at 11:43:26PM -0700, Rudy Setiawan enlightened us: I am trying to create backup servers out of the following specs: P4 3.0Ghz 1GB RAM RAID0 - 750GB x 2 (SATA) Hosts to backup: about 20 hosts with roughly 100GB each host. What will be the complications and the limitations? And also what are the recommendations? You realize that by using RAID 0, you are doubling the chance of losing all of your backups, because there is NO redundancy - so if either of your disks has a problem, you lose the entire array. If it is only temporary and you are then moving the data to something else, that might be all right, but I wouldn't feel comfortable with my backups (or anything other than scratch space, really) being on RAID 0. I would use RAID 1, or if you have disks to spare RAID 10. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: Configure error
On Wed, May 30, 2007 at 10:13:47AM -0500, Paul Crittenden enlightened us: I am trying to install amanda backup software on my server and I get the error, configure: error: *** --with-user=USER is missing, when I run ./configure. Can anyone point out what I need to do? Supply a --with-user=USERNAMEOFYOURCHOICE argument to configure. -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: problem with client on 64bit machine
On Wed, Mar 14, 2007 at 01:17:04PM -0500, Kenneth Kalan enlightened us: At 10:58 AM 3/14/2007, Jon LaBadie wrote: On Wed, Mar 14, 2007 at 10:29:49AM -0400, Greg Troxel wrote: Kenneth Kalan [EMAIL PROTECTED] writes: I've poke around the list and didn't find this problem there. I have a RHEL 4 box with two dual core Xeon's and 16 Gig's of ram. Naturally it's running x86_64 of Red Hat. It has the stock RH version of amanda, amanda-2.4.4p3-1 and amanda-client-2.4.4p3-1. This is a client box. The server is a 32bit box running the same versions (along with amanda-server-2.4.4p3-1). I cannot get the 64 bit box to backup. When I run amcheck it' replies with selfcheck request time out. Tried turning off the firewall, but no help. I install my boxes the same way, a script puts the server info into /var/lib/amanda/.amandahosts as well as putting the amanda file into xinetd.d. This works fine on all the 32 bit boxes, even ones configured after the 64 bit box. The RedHat RPM already puts the entries in xinetd.d - are you sure you haven't created a duplicate problem there? My experience is with NetBSD rather than RH, but I've used 64-bit machines with no problems. One setup has a i386 server with multiple clients, including sparc64 and works with krb4 auth. Another has a sparc64 server and i386 and sparc clients. So it's pretty unlikely there are serious 64-bit bugs in the code. Of course you could be having lib/lib64 confusion. I'd run ldd on the amandad binary and see if it links ok. Agreed. Does RH supply a real 64 bit compiled version. If not, i.e. it is a 32 bit version, it could be you have all the correct libaries in of the 64 bit variety but need to add some of the 32 bit ones. To answer an earlier question, nothing in /tmp/amanda, doesn't exist. RedHat moved it to /var/log where it probably should be anyway... The amanda files are compiled 64bit, they are in /usr/lib64/amanda and referencing the 64bit libraries. example: [EMAIL PROTECTED] ~]# ldd /usr/lib64/amanda/amandad libamclient-2.4.4p3.so = /usr/lib64/libamclient-2.4.4p3.so (0x002a9566c000) libamanda-2.4.4p3.so = /usr/lib64/libamanda-2.4.4p3.so (0x002a95773000) libm.so.6 = /lib64/tls/libm.so.6 (0x002a958ad000) libtermcap.so.2 = /lib64/libtermcap.so.2 (0x002a95a33000) libnsl.so.1 = /lib64/libnsl.so.1 (0x002a95b36000) libc.so.6 = /lib64/tls/libc.so.6 (0x002a95c4e000) /lib64/ld-linux-x86-64.so.2 (0x002a95556000) Just to double check once more, I've killed iptables on both boxes (server client), still not working. What does the output of chkconfig --list amanda chkconfig --list amandaidx chkconfig --list amidxtape give? Have you restarted xinetd once enabled? Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: disk-disk-tape best practices
On Thu, Mar 01, 2007 at 12:53:43PM -0500, Alastair Neil enlightened us: My experience with amanda in the past is that it will not write level 0 dumps only to holding space - is this incorrect? If this is the case then indeed my configuration is overly complicated. I have in the past maintained a 9 months backup history - that would be about 270 tapes - a lot of LTO2 media. I didn't mention but we are an academic environment so this starts to become cost prohibitive, as it is I was looking at only providing a semester ~ 3month backup window going forward. It most certainly does store level 0's to holding disk. I also am educational and have an LTO2 setup. My weekly backups are currently about 150GB, so I let them spool on a RAID1 holding disk for a week, then put a tape in and let them autoflush on the 7th day. I have a rotating set of tapes for 20 weeks, and do an archival dump quarterly. You probably just need to adjust your reserve parameter properly in amanda.conf. The default is 100% is reserved for degraded-mode backups. Set it to something smaller to allow fulls to fit. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: disk-disk-tape best practices
On Thu, Mar 01, 2007 at 04:10:33PM -0500, Alastair Neil enlightened us: On Thu, Mar 01, 2007 at 12:53:43PM -0500, Alastair Neil enlightened us: My experience with amanda in the past is that it will not write level 0 dumps only to holding space - is this incorrect? If this is the case then indeed my configuration is overly complicated. I have in the past maintained a 9 months backup history - that would be about 270 tapes - a lot of LTO2 media. I didn't mention but we are an academic environment so this starts to become cost prohibitive, as it is I was looking at only providing a semester ~ 3month backup window going forward. It most certainly does store level 0's to holding disk. I also am educational and have an LTO2 setup. My weekly backups are currently about 150GB, so I let them spool on a RAID1 holding disk for a week, then put a tape in and let them autoflush on the 7th day. I have a rotating set of tapes for 20 weeks, and do an archival dump quarterly. You probably just need to adjust your reserve parameter properly in amanda.conf. The default is 100% is reserved for degraded-mode backups. Set it to something smaller to allow fulls to fit. Matt Thanks for the pointer, I will try this. I assume I turn off autoflush? I'm going to try with a reserve of 25% - I have a 250 Gbyte holding disk. If you don't want all spooled dumps to flush to tape when one is inserted, then disable autoflush. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: AMANDA and gtar and Maildir and a couple of thousand mail users
On Fri, Dec 29, 2006 at 04:41:31PM -0500, Mark Hennessy enlightened us: I'm seeing this problem where gtar seems to take about 18 hours in total to back up. It's pulling the data off of a NetApp filer connected to the AMANDA server using Gigabit Ethernet. The total amount of data is like 120GB and I'm using a Quantum SDLT-220 drive. I'm not sure what could be causing this slowness in the backup. It seems to take like 6 hours to calculate the file list alone for this partition and the other drives/partitions that don't host Maildirs get have L0s done fine in like 2-3 hours at most so I would think it's because Maildir simply has so many files. Any good alternatives to gtar? Any commandline parameters to try to make the list creation be done more quickly or not done at all if possible? Thanks for any advice! You don't say what version of amanda you're using, but the estimate paramater of your dumptype should help considerably. I use CALCSIZE because I don't really care exactly how accurate the estimates are. There are a couple other options you should check out. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: 2.5.1p1 - Error connecting to windows clients
On Thu, Dec 07, 2006 at 02:20:40PM -0700, Ryan Castleberry enlightened us: I am running Debian testing. Upon initial upgrade to 2.5.1p1, amanda would not get ACKs from any machines. I ended up downloading a new amanda 2.5.1p1 debian package and reinstalling. I now have no problems getting to Linux machines, but am unable to reach windows machines via samba. As user backup, I am able to /usr/bin/smbclient //winpc/share. Any help would be appreciated. Thanks! Information that may be pertinent follows: $ amcheck -c daily Amanda Backup Client Hosts Check WARNING: winpc.mydomain.com: selfcheck request failed: timeout waiting for ACK Client check: 1 host checked in 30.013 seconds, 1 problem found (brought to you by Amanda 2.5.1p1) snip -- snip -- snip --- amandapass: --- //winpc/sharebackup%password - disklist: - winpc.mydomain.com//winpc/sharesamba Shouldn't that be someunixboxwithsamba.mydomain.org? Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: Need help backing up RHEL4 client with Debian tapehost
On Tue, Dec 05, 2006 at 11:07:56AM -0500, Zembower, Kevin enlightened us: I'm trying to backup my RHEL4 host 'www' using my Debian amanda server 'centernet.' Centernet backs up itself successfully, but can't back up the remote host www. I installed amanda-common, amanda-server and amanda-client from the Debian sarge packages on centernet, and amanda-client-2.4.4p3-1 and amanda-2.4.4p3-1 from the RHEL4 packages on www. The information at the end of this post shows the tcpdump of packets received on www when I do 'amcheck -c DBackup', which shows that they seem to be arriving at the host, but I get 'port amanda unreachable'. netstat doesn't show the amanda port open, but I'm not sure if it should (this is the only remote host I back up, so no compar4ison is possible). The xinet.d/amanda file seems correct; the backup host on www is 'amanda' and the backup host on the tapeserver centernet is 'backkup,' but I think I've got the correct answers in /var/lig/amanda/.amandahosts. The amanda debug files from the tapehost don't seem to show anything odd. Thanks for any help or suggestions in troubleshooting this problem. Did you run: # chkconfig amanda on # service xinetd restart ? Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: Need help backing up RHEL4 client with Debian tapehost
On Tue, Dec 05, 2006 at 11:50:52AM -0500, Zembower, Kevin enlightened us: Matt, you're amazing. I don't know which of the two commands did it, but it works now, at least minimally. I didn't know about the chkconfig command; I'll have to investigate it. It configures which services start on what runlevel. It also conveniently manages xinetd resources. I'm guessing there was a disable = yes; in your amanda xinetd script. I HUPped xinet with 'kill HUP pid-of-xinet' and also '/etc/init.d/xinet reload', but that might not have done the same as 'service xinetd restart' You did it the hard way :-) service foo args is equivalent to /etc/init.d/foo args Less typing involved, generally. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: Cron troubles, or config problem?
On Mon, Nov 27, 2006 at 07:31:25PM -0500, Gene Heskett enlightened us: Greetings; From the /var/log/cron file: Nov 27 19:20:01 coyote crond[7220]: Authentication service cannot retrieve authentication info Nov 27 19:20:01 coyote crond[7220]: CRON (amanda) ERROR: failed to open PAM security session: Success Nov 27 19:20:01 coyote crond[7220]: CRON (amanda) ERROR: cannot set security context An ls -l of the /home/amanda/.amandahosts file rw--- 1 amanda disk 413 Sep 19 10:47 /home/amanda/.amandahosts And its contents: coyote.coyote.den amanda amdump amindexd amidxtaped coyote amanda amdump amindexd amidxtaped coyote.coyote.den root amdump amindexd amidxtaped # gene.coyote.den root amdump amindexd amidxtaped # gene root amdump amindexd amidxtaped shop.coyote.den root amdump amindexd amidxtaped shoproot amdump amindexd amidxtaped shop.coyote.den amanda amdump amindexd amidxtaped shopamanda amdump amindexd amidxtaped So that *looks* ok. gene is powered down ATM. Now I recall there was something I had to do when switching to the newer tcp authentication model, so can someone please show me a list of files involved, and their perms? Or a url that shows this? Thanks. Cannot set security context points to selinux. You might do some ls -Z's and check the logs for errors in your selinux config. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: Release of amanda-2.5.1p2
On Fri, Nov 10, 2006 at 07:19:55AM -0800, Paddy Sreenivasan enlightened us: Giuseppe, Fabio, We will build 64bit versions of the RPMs and make them available soon. Can you please provide information on why you need 64 bit version of Amanda? On a 64-bit OS, it elliminates the need for bringing in all sorts of 32-bit libraries to run a single application. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: amflush during amdump
On Fri, Oct 27, 2006 at 01:10:57PM -0500, Lee, Raymond enlightened us: Will there be a conflict if I run amflush while an amdump is running? I'll be running amdump like this so that I can accumulate enough multiple dumps in my holding disk before I amflush it to tape: amdump DailySet1 -o tapedev=/dev/null -o tpchanger= (thanks to Jean-Louis for the taper options patch) Then I'll have a cron job that periodically checks the amount of space used in the holding disk. If the size is close to the capacity of one of my tapes, then amflush will run. So my question is, if amflush runs while amdump is running, will it try to flush the dump that amdump is currently working on and cause any problems? It won't run, it will complain there is another process already running. What I would do is set autoflush=yes in amanda.conf, then instead of running a separate cron job just write a wrapper around your amdump job. If the data is less than your required level, call it with your -o options, if it is above, call as usual and it will flush everything (including that run) to tape. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: labelling of reused tapes
On Wed, Oct 11, 2006 at 10:28:08AM -0400, Steven Settlemyre enlightened us: None of my tapes are new. I put the next 3 tapes that I have in the changer and run amcheck. It seems to be ok with 2 of the 3 tapes but gives me not an amanda tape (Input/output error) for slot 3. So I read the archives and they say to label the tape, so I run amlabel: BASH/usr/sbin/amlabel -f monthly Monthly53 slot 3 changer: got exit: 0 str: 3 /dev/nst0 labeling tape in slot 3 (/dev/nst0): rewinding amlabel: tape_rewind: rewinding tape: /dev/nst0: Input/output error amlabel: tape_rewind: rewinding tape: /dev/nst0: Input/output error amlabel: pid 17508 finish time Wed Oct 11 09:40:36 2006 Is this a tape problem or am I approaching this incorrectly? Looks like a tape error to me. Can you read/write it with standard tools (dd, tar, etc.)? Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: labelling of reused tapes
On Tue, Oct 10, 2006 at 03:00:32PM -0400, Steven Settlemyre enlightened us: I have a monthly backup config with runtapes=3 and 59 tapes in the tapecycle. The instructions I was originally given say to run a script that looks in tapelist, find the highest # then label the next 3 tapes starting from there. This seems wrong because then it wouldn't be a cycle. My question is when I do amcheck, it tells me (expecting tape Monthly04 or a new tape). But the next tape I have in my hand is currently labeled Monthly50 which is correct because it is in the tapelist as 20051107 Monthly50 reuse. How do I solve this problem? How do I tell amanda to look for Monthly50 instead of Monthly04? Or how do i have amanda use the tape I give it as long as it's not recently used? That's the Or New Tape part. If you've labelled a tape but not used it, it is considered a new tape. Amanda will also accept any tape that is more than tapecycle tapes old, so in my case my tapecycle is 16 tapes, but I actually use 20 in rotation, so that if one fails, I don't have to replace it immediately. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: backup/amflush oddity
On Wed, Sep 27, 2006 at 09:06:32AM -0400, Jeff Portwine enlightened us: Hoping somebody can clarify something about amanda for me. I have my backups set to run on a 3 day dump cycle with 6 tapes, running every evening except sunday night. It seems to work fine, we change the tape each morning so over the course of a week (if I understand how amanda works right) we get two sets of full backups. Sometimes we forget to change a tape, and most Saturdays we don't have anybody here to change the tape so on those days I thought that Amanda just backed up to the holding disk and then the next morning we could just change the tape, do an amflush, then change the tape again so that it would be ready for the next night. However, I've noticed that every time Amanda doesn't find a writeable tape (usually due to the tape not being changed that day) there is very little data written to the holding disk and the amflush is very very small. The day after this, the dump is much larger than usual compensating for this. When we change the tapes every day, each tape is usually around 50% full. On a day when we forget to change the tape or are unable to change the tape the amflush results in about 0.5% tape useage, and the day following the day we did the amflush tape useage is usually 85-90%. Is this how it's supposed to behave? How can I fix it to do a normal dump to the holding disk on days when either we can't change the tape or we forget to change the tape? Thanks for any explanation and/or tips. What settings do you have for your holding disk in amanda.conf? Specifically the reserve setting. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: amtapetype problems
On Fri, Sep 08, 2006 at 03:45:57PM -0400, Joshua Baker-LePain enlightened us: Please keep responses on the list. Also, top posting and not trimming posts are generally frowned upon. Thanks for pointing that out. I've had so many problems with this, I just assumed this to be another strange error and did not even notice the medium not present message duh. Also, I think load is not the right word to use since mt cannot load a tape, it must be done manually or by a robot. Precisely. Use 'mtx' to have the robot load the tape. After the tape drive goes through its (automatic) loading cycle, the tape will be ready for use. 'mt status' can confirm that. Furthermore, at least with my Overland robot, 'mtx unload' automatically does a 'mt offline' to eject the tape. So running 'mt' manually is pretty rare, really. Just for the record, I use mt's load command to pull a tape in that has been previously ejected, but not removed from the drive. After I dump to tape, I eject the tape, but don't physically remove it from the drive. That way if I need to restore a file from last week, it's as easy as mt -f /dev/nst0 load amrecover I eject the tape so that when amdump runs each night, it doesn't read the headers over and over, shortening the life of the tape (since I spool to holding disk all week and flush to tape once a week) and forces me to run another command before I can have a typo erase all of last weeks dumps :-) Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: amdump problem
On Thu, Aug 17, 2006 at 01:30:18PM +0200, Natalia García Nebot enlightened us: well, i have configured amanda only in my server host which is conected to a tape device. I have only one tape yet. I have put in my disklist thie line: aroprod /home/natalia always-full I execute this command: su amanda -c amdump DiariaPrueba In my log.20060817.2 file i have this: START driver date 20060817 DISK planner aroprod /home/natalia START planner date 20060817 WARNING planner tapecycle (1) = runspercycle (1) INFO planner Adding new disk aroprod:/home/natalia. START taper datestamp 20060817 label DiariaPruebaTape01 tape 0 ERROR planner Request to aroprod timed out. FINISH planner date 20060817 WARNING driver WARNING: got empty schedule from planner STATS driver startup time 30.000 INFO taper tape DiariaPruebaTape01 kb 0 fm 0 [OK] FINISH driver date 20060817 time 34.380 what am I doing wrong? Have you run amcheck? Better start there before trying to dump anything. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: amanda client inetd problem
On Fri, Aug 11, 2006 at 09:36:21AM -0400, Jeff Portwine enlightened us: amanda dgram udp wait backup /usr/local/libexec/amandad amandad Or was it nowait, and you changed it to wait (or fixed the username backup), but forgot to sig-HUP the inetd process after you fixed it? No, I added the inetd entry as listed above and then actually rebooted the machine to make sure inetd restarted properly. What happens when you execute the command /usr/local/libexec/amandad as user backup manually? $ /usr/local/libexec/amandad /usr/local/libexec/amandad: error in loading shared libraries: libamclient-2.5.0p2.so: cannot open shared object file: No such file or directory However, that library does exist... $ ls -l /usr/local/lib/libam* -rwxr-xr-x1 root root 361796 Aug 11 08:07 /usr/local/lib/libamanda-2.5.0p2.so -rw-r--r--1 root root 1270664 Aug 11 08:07 /usr/local/lib/libamanda.a -rwxr-xr-x1 root root 875 Aug 11 08:07 /usr/local/lib/libamanda.la lrwxrwxrwx1 root root 20 Aug 11 08:07 /usr/local/lib/libamanda.so - libamanda-2.5.0p2.so -rwxr-xr-x1 root root99610 Aug 11 08:07 /usr/local/lib/libamclient-2.5.0p2.so -rw-r--r--1 root root 239810 Aug 11 08:07 /usr/local/lib/libamclient.a -rwxr-xr-x1 root root 889 Aug 11 08:07 /usr/local/lib/libamclient.la lrwxrwxrwx1 root root 22 Aug 11 08:07 /usr/local/lib/libamclient.so - libamclient-2.5.0p2.so so I'm not sure what the problem is. Is /usr/local/lib in your ld.so.conf and did you run ldconfig after installing amanda? Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: can only run amanda tests as root
On Fri, Aug 04, 2006 at 07:03:39AM -0400, Gene Heskett enlightened us: This is a frequent problem when using a 'packaged' version of amanda. The rpm packagers in particular have demonstrated many times that they do not understand how amanda treats security issues. Amanda's build gives it enough perms to do the instant job each piece needs to do, and no more. This is why, when potential users run into these sort of problems, that we universally recommend that it be built from the tarball, following amanda's instructions so that the amanda build and install can be done correctly. Not to get off topic, and no offense Gene, but that second sentence is entirely inaccurate. There are several problems with Amanda RPMs which require compromises be made, but not understanding security and how amanda works is definitely *not* one of them. What some of the problems are include: - Inability to know the hostname of the amanda server (or any server for that matter) on the clients network. Since amanda requires these at build time, the only logical compromise is to use localhost - Inability to know what user the software will be built as (if rebuilding a source RPM). There are RPM mechanisms in place that fix that, though not pretty, which set ownership to the appropriate user and permissions. I would argue that the fact these are in there (at least in the RedHat-produced RPMs) counter your point about not understanding the security needs. As always, I consider http://www.math.ohiou.edu/~hyclak/casit/amanda/ a worthwhile read for anyone using an RPM-based system, and recommend that users of Amanda rebuild SRPMs for their environment. Jay has made it easier with the latest Fedora Development RPMs to do that in the future. Respectfully, Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: Amrecover help needed
On Thu, Jul 20, 2006 at 07:47:08PM +0100, Anne Wilson enlightened us: I'm not understanding amrecover, and need some help. Specifically, I want to recover a file foo.xls which originated from /home/anne/Documents/Spreadsheets. I don't want it to overwrite the existing copy, so want to recover it to a named directory. My vtapes are in /Backup/amandatapes/Dailys. I created a directory /Backup/Recover, then started amrestore. [EMAIL PROTECTED] ~]# amrecover AMRECOVER Version 2.5.0-20060323. Contacting server on borg ... 220 borg AMANDA index server (2.5.0-20060323) ready. 200 Access OK Setting restore date to today (2006-07-20) 200 Working date set to 2006-07-20. Scanning /tmp/dumps... 200 Config set to Daily. 200 Dump host set to borg. Trying disk / ... Trying disk rootfs ... Can't determine disk and mount point from $CWD '/root' amrecover cd .. Must select disk before changing directory amrecover exit 200 Good bye. Just do a setdisk when you get the prompt to change to the appropriate DLE. 200 Disk set to /. /root/restore WARNING: not on root of selected filesystem, check man-page! amrecover setdisk /export 200 Disk set to /export. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: Amrecover help needed
On Thu, Jul 20, 2006 at 09:15:55PM +0100, Anne Wilson enlightened us: 'man amrecover' and specifically look at the 'setdisk' command. Depending on what your DLE looks like, you probably need to 'setdisk /home'. Right - /home is the root directory of one DLE, and the required file is in /home/anne/Documents/Spreadsheets. According to the examples in man:amrecover I should now be able to ls the file, but 200 Disk set to /home. amrecover ls -l bankrec.xls 2006-07-20 nigel/ 2006-07-20 micky/ 2006-07-20 lost+found/ 2006-07-20 gillian/ 2006-07-20 david/ 2006-07-20 anne/ 2006-07-20 andy/ 2006-07-20 amanda/ 2006-07-17 amanda.exclude~ 2006-07-17 amanda.exclude 2006-07-20 .Trash-0/ 2006-07-20 . Invalid command - syntax error amrecover setdisk /home/anne/Documents 501 Disk borg:/home/anne/Documents is not in your disklist. amrecover setdisk /home 200 Disk set to /home. amrecover What is the next step? Anne The command environment within amrecover is pretty limited. There is no -l option (or any option, for that matter) to ls. What you need to do would be something like: setdisk /home cd anne cd Documents cd Spreadsheets add bankrec.xls extract Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: Restoring from tape when Amanda server failed
On Wed, Jul 19, 2006 at 02:07:42PM -0700, gil naveh enlightened us: Thanks for all the help, but I have a problem to restore the files. When I type: root@ # dd if=/dev/rmt/0n ibs=64k | ufsrestore if - I recieve the following error message: read: Invalid argument 0+0 records in 0+0 records out Volume is not in dump format But as far as I know it should be in a dump format!!! because in the Amanda.conf I defined the backup as: You forgot to strip off the amanda header at the beginning of the file. Usually this is dd if=/dev/tape bs=32k skip=1 -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: Port NNNN not secure (revisited)
On Tue, Jul 18, 2006 at 10:53:52AM -0700, Mike Allen enlightened us: Jon LaBadie wrote: On Mon, Jul 17, 2006 at 09:32:01AM -0700, Mike Allen wrote: The following occurred while compiling with the values you suggested: snip make install-data-hook chown operator /usr/local/man/man8/amanda.8 chgrp operator /usr/local/man/man8/amanda.8 chown operator /usr/local/man/man8/amanda.conf.5 chown: /usr/local/man/man8/amanda.conf.5: No such file or directory *** Error code 1 Stop in /usr/ports/misc/amanda-client/work/amanda-2.4.5p1/man. *** Error code 1 Stop in /usr/ports/misc/amanda-client/work/amanda-2.4.5p1/man. *** Error code 1 Stop in /usr/ports/misc/amanda-client/work/amanda-2.4.5p1/man. *** Error code 1 Stop in /usr/ports/misc/amanda-client/work/amanda-2.4.5p1. Why is it looking in /usr/local/man/man8 for a man5 manual page? I have seen this before. That was a defect created when the man page for amanda.conf(5) was split from amanda(8). It was corrected in a later release. As I recall it was a very simple editing correction to the makefile(s) to do the installation. Otherwise manually install the manpages. I have done the following: tcpportrange = 5,50100 udpportrange=512,1023 Did a 'make distclean' before configuring with the above parameters. Did the same configuring on both the client and server. Opened the specified ports on my Netgear firewall (ie. Port-forwarded the ports to the IP address of my tape server host) The results seem to be the same: 09:07:09:07:45.431266 IP famrad.familyradio.org.50327 familyserv.familyradio.org.amanda: UDP, length: 123 09:07:45.475830 IP familyserv.familyradio.org.amanda famrad.familyradio.org.50327: UDP, length: 50 09:07:45.476738 IP familyserv.familyradio.org.amanda famrad.familyradio.org.50327: UDP, length: 109 09:07:45.480662 IP famrad.familyradio.org.50327 familyserv.familyradio.org.amanda: UDP, length: 50 Why is port 50327 being referenced? I don't have any ports opened there. What am I doing wrong? Maybe I have to change to an IP-Tables based firewall? Does anybody have any experience or insight on this? If you're going through a NAT, then there is a port translation that is taking place. That's probably where that's coming from. Can you lock specific ports on the firewall device to not be remapped? If not, you'll probably want to look into some sort of VPN or tunnel. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: Difference between the amrestore and amrecover
On Mon, Jul 17, 2006 at 05:57:45AM -0700, silpa kala enlightened us: I need some clarification. I worked on amrestore , which we can retrieve the backup image from holding disk or tape. amrestore /dev/nst0 amrestore holdingdiskname But I didn't understand the amrecover. I feel amrecover is also developed for the same purpose. Why we need two commands for the restore? Is there any detailed technical mannual is available for amrecover? Please clarify the doubts. amrecover gives you an interactive interface to select specific files. amrestore is used for pulling entire DLE's from tape. There is a man page for amrecover and source code. What more do you need? Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: making xfs dump work
On Thu, Jul 13, 2006 at 04:30:01PM -0400, Julian C. Dunn enlightened us: On Thu, 13 Jul 2006 at 3:57pm, Julian C. Dunn wrote I have /usr/sbin/xfsdump already but Amanda doesn't seem to know how to call it, even though sendsize is aware that it's an XFS filesystem: Was xfsdump there when you *compiled* amanda? That's when the locations of such things get noted. Ah... I'm using the binary vendor RPM which I assume doesn't contain XFS support. Time to go compile my own. I'd start here: http://www.math.ohiou.edu/~hyclak/casit/amanda/ -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: Problems using amanda to backup SMB shares
On Tue, Jul 11, 2006 at 11:30:23AM -0400, Nathan Weston enlightened us: I'm a first-time amanda user, trying to backup a linux machine and a few windows shares. I went through the tutorial at amanda.zmanda.com/quick-backup-setup.html, modifying things to fit my setup. The linux machine seems to be ok, but I'm getting some errors with the SMB share. The windows machine in question is lens, and the backup server is penguin2. On penguin2, I have a disklist entry like this: penguin2.genartsi.com //lens/ccomp-user-tar When I run amcheck, I get the following errors: Amanda Backup Client Hosts Check ERROR: penguin2.genartsi.com: [samba access error: //lens/c: Error connecting to 192.168.0.43 (Permission denied): Connection to lens failed: returned 1] ERROR: penguin2.genartsi.com: [can not read/write /etc/amandates: Permission denied] Client check: 2 hosts checked in 0.136 seconds, 2 problems found Looking in selfcheck.debug, I see that it's running smbclient like this: smbclient \\lens\c -U amanda -E -c quit When I run the same command manually, it prompts for a password, and when I enter that it connects fine. But I don't see how amanda is giving this password to smbclient (since it obviously can't prompt for it). Maybe that's the problem? Yes, you need to create an /etc/amandapass file with the appropriate username/password combinations. man amanda for the details. As for the second problem, I don't see anything in the .debug files that gives me a clue. The permissions for /etc/amandates look like this: -rw-r- 1 amandabackup disk 0 Jul 10 17:20 /etc/amandates Any ideas? Is the user you configured amanda as actually amandabackup or is it really just amanda? Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: Problems using amanda to backup SMB shares
On Tue, Jul 11, 2006 at 12:08:27PM -0400, Nathan Weston enlightened us: Matt Hyclak wrote: Yes, you need to create an /etc/amandapass file with the appropriate username/password combinations. man amanda for the details. As for the second problem, I don't see anything in the .debug files that gives me a clue. The permissions for /etc/amandates look like this: -rw-r- 1 amandabackup disk 0 Jul 10 17:20 /etc/amandates Any ideas? Is the user you configured amanda as actually amandabackup or is it really just amanda? I already created /etc/amandpass, and I really am running amcheck as the user amandabackup -- I installed from the zmanda RPMs and this is how they set things up. But, I seem to have found the source of the problem -- I'm running Fedora Core 4 which has SELinux enabled by default. When I turn that off, amcheck runs with no errors. Is there an easy way to make amanda play nice with SELinux? If not, I can live with SELinux disabled, since our backup server is only accessible from the LAN anyway. Ah, that can do it, too. I would do an ls -Z on /etc/amandapass to see what the security context is. You can also use audit2allow to see what context the amandabackup user is permitted to use and adjust accordingly. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: RPM for RHEL3
On Mon, Jul 03, 2006 at 11:56:23AM -0500, Gordon J. Mills III enlightened us: Does anyone have an amanda rpm for RHEL3? I have a client with this server and I would like to put amanda on it to do the backups. Will one of the rpm's for rhe[ls]4 work? I would prefer to use a package instead of compiling. I tried to compile and apparently the server doesn't have all the necessary packages to compile. TIA, Gordon I recommend reading this: http://www.math.ohiou.edu/~hyclak/casit/amanda/ Recompiling is really the best way to do things, the page above shows you how to do that within the RPM environment so you get the advantages of both. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: Info on backing up windows servers via amanda
On Thu, Jun 29, 2006 at 01:24:45PM -0400, Lengyel, Florian enlightened us: I'm running amanda version 2.4.4 under Cent OS. So far it has been used to back up Linux machines; now I've acquired four windows 2003 machines that I'd like to back up with Samba. Does someone on the list have experience with this? I'm sure several people do, and you'll probably get better answers if you ask a better question. Are you having a problem with this? Search the archives, there are many best-practice types of threads regarding this subject. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: Info on backing up windows servers via amanda
On Thu, Jun 29, 2006 at 02:10:43PM -0400, Lengyel, Florian enlightened us: I'm not having a problem with backing up a windows system to a linux amanda server; there seems to be trouble interpreting a request for info or suggestions on where to find it, though. Is this a better question: where do I find out about the various methods people use? It's pretty well documented in the amanda docs. You've basically got a few options: 1. smbclient from some *nix machine to the windows box 2. smbmount the windows box on a *nix machine 3. Install cygwin on windows and do a native amanda compile 4. Use native windows tools to dump to files in a directory being backed up I'm sure all of these have been covered in the mailing list, and the documentation, so it's up to you which best suits your needs. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: Appropriate number of tapes in a set
On Mon, Jun 26, 2006 at 04:14:29AM -0700, Joe Donner (sent by Nabble.com) enlightened us: Hi, can anyone please help me understand this: Apparently, you have to calculate how many tapes you need for your backup schedule as the number of tapes needed for every day the backup will run in your dumpcycle + one extra tape to avoid the risk of overwriting a full backup performed at the beginning of the previous cycle (quoted very loosely). The minimum amanda will run with is runspercycle * runtapes. I think most people recommend 2 * runspercycle * runtapes + 1, so that there are 2 complete sets of backups, plus a spare tape. I want to: 1. Run a backup job once a day Monday to Friday (all my data will fit onto one tape), All data Monday-Friday will fit on one tape, or each night all data will fit on one tape? 2. Have four weeks' worth of historical backups, i.e. be able to restore data to any given point within the last four weeks, and 3. Permanently archive one monthly full backup. So how would I calculate the appropriate number of tapes needed? 5 days * 4 weeks + 1 = 21 tapes ?? Assuming 1. above means you need just 1 tape per night, then for your 4 weeks of historical backup, 21 tapes would be sufficient. If I don't force full backups, and therefore full backups are sort of scattered between different tapes, then I'll have to take out multiple tapes every month for archiving, right? How should that be managed? You have 2 choices: Remove an entire set of tapes (1 complete dumpcycle, in your case 5 tapes), or run a second configuration with its own set of tapes that forces fulls. I do the latter. What would appropriate values be for dumpcycle, runspercycle, and tapecycle? I'm thinking: dumpcycle 7 days (to ensure a full backup at least once a week) runspercycle 5 days (Monday to Friday every week) tapecycle 21 (number of tapes in rotation assuming my calculation above is correct) That would work nicely. I'd really appreciate your advice on this type of backup scheme, because I'm struggling a bit to understand Amanda - they should have called her George or something :) You're doing just fine :-) Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: Appropriate number of tapes in a set
On Mon, Jun 26, 2006 at 05:01:44AM -0700, Joe Donner (sent by Nabble.com) enlightened us: Thanks very much for clarifying that to me. Yes, I meant that (at least at present) all data will fit onto one tape for every day of the week, i.e. if a full backup is done every day Monday to Friday, then Monday's full backup will fit onto one tape, Tuesday's will fit onto one tape, and so on. About the second configuration for doing monthly archival backups: 1. I assume then that you'd have to do something once a month to prevent the normal Amanda backup job from running while the monthly full one is active? I seem to remember I read somewhere about placing a file called hold inside an Amanda directory to cause the normal daily run to pause so it doesn't interfere with your monthly job - and so that you don't have to modify your cron job each time? Yes, I usually put the Monthly tape in the drive, then schedule an at job for later in the evening that basically does touch /etc/amanda/Dailies/hold amdump Monthly rm /etc/amanda/Dailies/hold 2. In other words, if you have a second configuration for monthly archives, then you really just happily rotate your 21 tapes for the normal daily backups, and then replace them one at a time either when they die naturally or when they've reached their expiry date? Thanks again - I think this has made things click somewhere in my head. Yep, that's it. With the Monthly archive configuration, you'll probably want to set record to no in your amanda.conf so that your system doesn't count those as being full dumps and messing up your schedule every month. Other than that, it's pretty straightforward. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: amcheck reading label: Device or resource busy
On Sat, Jun 24, 2006 at 11:18:06PM -0700, Robert Grey enlightened us: On Jun 24, 2006, at 6:27 PM, Jon LaBadie wrote: On Sat, Jun 24, 2006 at 05:46:27PM -0700, Robert Grey wrote: I am setting up Amanda on a Gentoo box with an Overland LoaderXPress and can't seem to get past an amcheck issue. The amdump operation works fine but amcheck gives this error: amdump is working but amcheck is not? I'd first check permissions on amcheck. It should be root owned and setuid'ed. setuid is set and the owner is root; this was set by the gentoo emerge (build) process: dev64 daily # ls -la /usr/sbin/amcheck -rws--x--- 1 root amanda 40640 Jun 23 00:08 /usr/sbin/amcheck I am fairly sure the permissions are correct because amcheck running as amanda does switch the tapes after it can't work with the drive: dev64 ~ # ls -la /dev/sg0 crw-rw 1 root tape 21, 0 Jun 23 00:43 /dev/sg0 dev64 ~ # ls -la /dev/changer lrwxrwxrwx 1 root root 3 Jun 24 12:07 /dev/changer - sg0 dev64 ~ # ls -la /dev/nst0 crw-rw 1 root tape 9, 128 Jun 23 00:43 /dev/nst0 dev64 ~ # ls -la /dev/tape lrwxrwxrwx 1 root root 4 Jun 24 12:07 /dev/tape - nst0 dev64 ~ # sudo -u amanda id uid=87(amanda) gid=87(amanda) groups=26(tape),87(amanda) --- Here is my amanda.conf: --- org daily mailto removed # space separated list of operators at your site dumpuser amanda # the user to run dumps under is amanda user or amanda's group allowed to read the tape device? see above tape is the group which can read/write the tape drive. amanda is part of that group, but it is not it's primary group. I didn't see your xinetd config file, but do you have groups = yes set, so that xinetd will allow the permissions of secondary groups to be used? Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: No Level 0 backup possible
] FINISH planner date 20060617 time 3.387 START taper datestamp 20060617 label Dailys-6 tape 0 SUCCESS dumper borg /home 20060617 5 [sec 19.738 kb 19057 kps 965.5 orig-kb 104890] SUCCESS chunker borg /home 20060617 5 [sec 19.696 kb 19057 kps 969.2] SUCCESS taper borg /home 20060617 5 [sec 0.237 kb 19104 kps 80575.3 {wr: writers 597 rdwait 0.000 wrwait 0.147 filemark 0.000}] SUCCESS dumper borg /home/anne/Photos 20060617 1 [sec 0.386 kb 3700 kps 9577.0 orig-kb 3700] SUCCESS chunker borg /home/anne/Photos 20060617 1 [sec 0.389 kb 3700 kps 9591.2] SUCCESS taper borg /home/anne/Photos 20060617 1 [sec 0.040 kb 3744 kps 91576.2 {wr: writers 117 rdwait 0.000 wrwait 0.025 filemark 0.000}] SUCCESS dumper borg /Public 20060617 1 [sec 0.050 kb 8 kps 157.1 orig-kb 80] SUCCESS chunker borg /Public 20060617 1 [sec 0.053 kb 8 kps 745.3] SUCCESS taper borg /Public 20060617 1 [sec 0.008 kb 64 kps 7702.5 {wr: writers 2 rdwait 0.000 wrwait 0.000 filemark 0.000}] INFO taper tape Dailys-6 kb 22912 fm 3 [OK] FINISH driver date 20060617 time 38.231 It looks to me as though it is not attempting to span the vtapes. Can someone please point me at the likely cause? I did have spanning working at one time, so I must have inadvertently changed something vital. You could check how your settings are for each DLE individually and make sure they seem reasonable (what you expect). For example: amadmin Dailys disklist borg /home amadmin Dailys disklist borg /home/anne/Photos Looking again at the wiki page re splitting dumps, I have created a new dumptype: define dumptype user-tar-span { root-tar tape_splitsize 400Mb comment tape-spanning user partitions dumped with tar } The wiki then goes on to say Now, in your disklist file, just invoke that dumptype: nutdumpling /export/home/building/amanda/dumpdata user-tar-span This format doesn't make any sense in terms of my disklist, which for this DLE reads borg /home { program GNUTAR compress CLIENT BEST index yes exclude list amanda.exclude } 1 What would be the syntax needed to use user-tar-span here? Sorry of this sounds naive but I'm still missing some basic concepts :-) You can replace the program GNUTAR line with a dumptype. In your case, you would change your DLE to be borg /home { user-tar-span compress CLIENT BEST # This may already be defined in root-tar index yes# This may already be defined in root-tar exclude list amanda.exclude } 1 Your compress and index lines are overriding anything in root-tar. It may be redundant, but I just thought I'd point it out. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: tar's default block size shoe-shinning
On Mon, Jun 19, 2006 at 02:59:16PM +0200, Cyrille Bollu enlightened us: indeed... Thank you for your answers Joshua, I will try a dd with bs=2048k. Then probably I will compile amanda... Bouh I don't like to do that in a production environment Understandable. I have some tips at http://www.math.ohiou.edu/~hyclak/casit/amanda/ for using Amanda on RPM based systems. I recommend you rebuild the SRPM anyway, to get rid of necessary but bad default compilation settings. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: No Level 0 backup possible
On Mon, Jun 19, 2006 at 02:26:49PM +0100, Anne Wilson enlightened us: You can replace the program GNUTAR line with a dumptype. In your case, you would change your DLE to be borg /home { user-tar-span compress CLIENT BEST # This may already be defined in root-tar index yes# This may already be defined in root-tar exclude list amanda.exclude } 1 Your compress and index lines are overriding anything in root-tar. It may be redundant, but I just thought I'd point it out. I'm sure that the problems are being caused by conflicting/overriding commands - it's just finding them that's the problem :-) I tried commenting out the compress and index lines, but attempting to run the dump still tells me that Line 271 (root-tar) needs a dump-type parameter. Do you have root-tar defined as a dumptype somewhere in amanda.conf? I has to be defined before you can reference it. It will start define dumptype root-tar { global program GNUTAR ... } Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: Is tape spanning documented anywhere?
On Mon, Jun 12, 2006 at 10:18:46AM +0200, Toralf Lund enlightened us: I haven't been following the posts to this list too closely, or bothered to upgrade amanda, for some time (since our existing setup *works*...), so I didn't find out until right now that tape spanning is supported in the current release. Anyhow, I'd really like to know more about how the spanning actually works. Is it documented anywhere? http://www.amanda.org/docs and FAQ still say that the option does not exist... Try http://wiki.zmanda.org/index.php/Splitting_dumps_across_tapes Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: tapetype definitions
On Fri, Jun 02, 2006 at 07:08:39AM -0700, Joe Donner (sent by Nabble.com) enlightened us: Dear all, I've just run amtapetype -f /dev/nst0 for Amanda to generate a tapetype definition for an HP Storageworks SDLT 320 tape drive. The results came back as: define tapetype unknown-tapetype { comment just produced by tapetype prog (hardware compression on) length 135040 mbytes filemark 39 kbytes speed 2272 kps } amtapetype did slightly complain about hardware compression being enabled, but I couldn't see a way of disabling that. Would you, in your experience, think that this tapetype definition seems accurate and acceptable to use for real backups? This is the last outstanding piece of the Amanda puzzle before putting Amanda into operation. The tape drive's capacity is 160Gb uncompressed and 320Gb compressed, according to its specifications. So I'm not too sure about the length 135040 mbytes mentioned in the generated tapetype. Won't this cause Amanda to not use the drive's full capacity? Any advice will be much appreciated. Yes, disable hardware compression. The reason amanda only detected about 135GB of your 160GB tape is because compressing already-compressed data usually causes it to expand. The only exception I've seen is LTO, which is smart enough to leave it alone. Disabling hardware compression happens at the OS or tape drive level. You can use mt to turn it of on most Unix-like systems, or better yet disable it with a dip switch/jumper on the drive itself. 2nd suggestion: forget the number 320GB. It is purely marketing. Your tapes will hold 160GB of data. The data may be compressed, and you may be able to fit 320GB on there, you may be able to fit 400GB on there, or you may only be able to fit 162GB on there. It all depends on your data and how compressable it is. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: tapetype definitions
On Fri, Jun 02, 2006 at 08:14:28AM -0700, Joe Donner (sent by Nabble.com) enlightened us: Firstly - thank you for the replies. Secondly - I'm now more confused than before! Is my understanding correct then that I can conceivably use this tapetype definition, but that I probably shouldn't expect to be backing up more than approximately 160GB to a tape? Or, more to the point, approximately 135GB in this case? And that I should really disable compression (but that this isn't necessarily required) and accept the tape's capacity of around 160GB? I'm new to Linux (Redhat ES 3 in our case) and Amanda, so wouldn't mind things being spelled out to me - if you wouldn't mind doing that. Not at all. 1. Tapes for your tape drive hold 160GB of data. Period. 2. When you try to compress data that is already compressed, it expands. This is evidenced by the fact that amanda was only able to write 135GB of compressed data to your tape drive: when it got through the tape compressor, it expanded to 160GB. 3. Amanda works best when using software compression, because it knows how many bytes were sent to the drive are how many will be written to tape. If you wish to use hardware compression, you have to lie to amanda to tell it what tape size you think you'll have. SDLT's Marketing Department would have you believe that is 320GB. The real answer is it depends on what you're backing up. If it's directories of .jpg images and tarballs, you won't get anywhere near 320GB. If it's a bunch of text files, you might even exceed 320GB. The fact of the matter is it's a guessing game with hardware compression. Using software compression, amanda will compress your data before sending it to the tape drive, so it knows how to fit 160GB on the tape. That 160GB, when expanded, could be anywhere from 160GB - 300GB or more. Again, depends on your data. 4. You can use the tapetype as is, and amanda will base logic decisions on that information, but once in backup mode, amanda writes to tape until it hits an error (generally End Of Tape (EOT), but sometimes there are media errors, etc.). Does that help? Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: Rif: RE: Cluster backup
On Wed, May 31, 2006 at 03:37:44PM +0200, Nicola Mauri enlightened us: Paul, thanks for your suggestions. On our machines, all network masks appear to be consistent. We'll try to configure different client instance on xinetd. Another frequent message error I didn't mention in my first post is: aborted:nak error: amandad busy which may sometimes lead to: driver: (aborted:nak error: amandad busy)(too many dumper retry) virutalA /apps/a RESULTS MISSING Any ideas? If I had to guess, I'd say that you had for example Server1 which has an IP and also the VirtualA IP. In you disklist you have DLEs for both Server1 and VirtualA. Amanda contacts Server1 to start an estimate/dump. Amanda tries to contact VirtualA to do the same, but amandad is already running on that machine. Sound like a possibility? Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: Backup with dump program
On Thu, May 18, 2006 at 07:21:42PM +0530, Rajeh kuri enlightened us: I am new to the Amanda utility and evaluating it on SUSE Linux 10(2.6.16 kernel) I have been trying to backup the my home directory using dump program. The amstatus showing planner: [backup path,all estimate failed] but iam able to backup using GNU TAR program. Can I know what is going wrong. If iam not giving enough info for clue please let me know what need to be sent. Generally dump is used to do complete filesystems, not subdirectories. You *can* use dump for subdirectories, but you can only do full (level 0) dumps, not incrementals. Without more information, we can't tell you what's wrong. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: Can't get a Level 0 backup
On Wed, May 10, 2006 at 09:16:22PM +0100, Anne Wilson enlightened us: And there is the likely culprit, exclude file vs. exclude list. I.e. you are excluding any file matching the single pattern amanda.exclude. What you want is to exclude any file matching any pattern in the list of patterns in amanda.exclude. So I need to replace the line 'exclude file amanda.exclude' with exclude list amanda.exclude? OK - I'll try that. If it works it should get me where I wanted to be. Exactly. I even do exclude list optional .amanda.exclude so if I don't need to exclude anything from a DLE, I don't have to create the empty file. I'd still like to understand, though, why it thought there was not enough room. I would have thought that there was plenty, even for a full backup? Even though there's space on the disk, how big are your vtapes set to be? I'm guessing less than your directory size. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: How to amverify
On Thu, May 04, 2006 at 08:51:07AM +0700, Olivier Nicole enlightened us: How to tell amverify to stop if there is no tape in the tape drive? Every night I have a cron that does amdump then amverify. If I forgot to put a tape in the drive, the dumps go to the holding disk and amdump ends nicely. But amverify will continue running until I insert a tape next morning. And when I want to flush my holding disk, i find out that amverify is verifying the tape I just inserted and it messes up everything. So it would be nice that amverify could be told to abort if there is no tape in the drive. Is there a way to do so? I would just script it: mt -f /dev/tape status | grep ONLINE /dev/null if [ $? -eq 0 ]; then amverify fi Flavor to taste. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: Defining a sensible backup routine
On Mon, Apr 17, 2006 at 12:15:07PM +0100, Anne Wilson enlightened us: I did wonder if it was possible to define a completely separate set of parameters for these video directories, using a monthly rotation, say, but I get the impression that that's not possible. You could do a second config for this. I currently run 2 configs, one is a Daily backup, the other is a Quarterly archival dump that uses the same disklist, but doesn't record anything and forces fulls. I don't have the second config in cron anywhere, when it gets to the right time I simply touch /etc/amanda/Dailies/hold to prevent the Dailies from running, then run amdump Fulls by hand. When it's done, I remove the hold file and amanda continues with the Dailies as usual. (Usually this is one command: amdump Fulls rm /etc/amanda/Dailies/hold, so I don't have to babysit). Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: recover disks under amanda
On Tue, Apr 11, 2006 at 03:04:46PM -0300, Pedro Drimel Neto enlightened us: I'm using amrecover to recover the files of disks that amanda copy to the TAPE. I have a simple doubt: When I do a amdump it record the label BACKUP001 it OK but if I try to recover the tape BACKUP002 after the dump of BACKUP001 it want only the BACKUP001, how do I say to amrecover that I want to recover data files of BACKUP002 ? Any idea ? PS.: If I have to do this recover with another tool, no problems. Best regards. amrecover restores based on date. It will tell you which tapes it needs to return you to that date. amrestore will do mass dumping of backups from whichever tape you chose. That may be more what you're looking for. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: Amanda daily report question
On Fri, Apr 07, 2006 at 01:10:40PM +0100, Anne Wilson enlightened us: I'm still at the stage of trying different settings and running test dumps. One such run has just completed, with the following report: snip -- snip -- snip FAILURE AND STRANGE DUMP SUMMARY: borg /home lev 1 STRANGE snip -- snip -- snip I guess the INBOX warnings are where a message has been deleted while amanda was running. KMail will not normally be open when amanda is scheduled to run, so that's not a problem. Is there anything else there that is worrying? Strange is OK, it's the Failure's you have to worry about. Looks good to me. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: Amanda daily report question
On Fri, Apr 07, 2006 at 01:29:58PM +0100, Anne Wilson enlightened us: Strange is OK, it's the Failure's you have to worry about. Looks good to me. Thanks, Matt. I think that only leaves splitting down the big one into smaller jobs, and I'm ready to schedule. Basically, I have some big directories containing video and photos, that don't compress well, but don't change very much either, while the rest is small files in formats that should compress well but are more fluid. I'm guessing that it would be useful to split them to reflect that. Or doesn't it matter very much? If you know they're not compressable, you could skip that for those specific DLE's (e.g. use user-tar instead of comp-tar). That should save some time and CPU cycles on your backup cycle. If you wanted to get really fancy, you could probably change the frequency of Level 0's on the mostly-static DLEs to save space, but I'm not sure I'd bother with it. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: gnu tar-1.15+ for RHEL4
On Fri, Apr 07, 2006 at 01:17:20PM -0400, Jay Fenlason enlightened us: i am installaing amanda on a RHEL4 workstation. when using amanda-backup_client-2.5.0-1.rhel4.i386.rpm from the zmanda site, it fails with: error: Failed dependencies: tar = 1.15 is needed by amanda-backup_client-2.5.0-1.rhel4.i386 yet when i have searched for a tar = 1.15 rpm i have only found rpms for RHEL4 going up through tar-1.14-9.RHEL4.src.rpm (this on rpmfind). = 1.15 rpms are only available for SuSE, Mandriva, Fedora and Mandrake. is it possible to use an rpm from one of these other distros? what do other folks do about this? You could always grab the tar .src.rpm from Rawhide or Fedora Core 5 and rebuild it on your RHEL-4 box. Or you could grab the amanda .src.rpm, rewrite the tar dependency, and rebuild it. That's probably what I would do. It sounds to me like a packaging error that the rpm doesn't work with a version of tar that comes with the OS and is known to have been patched to work with Amanda. But what do I know about packaging? :-) Maybe if *someone* would finish packaging 2.5.0 for FC, my instructions at http://www.math.ohiou.edu/~hyclak/casit/amanda/ would be applicable. For the zmanda packagers: The RHEL4 packages really shouldn't require tar = 1.15. RHEL 4 will always have tar 1.14 with backported patches. Perhaps not the best idea, but that's their policy, so we have to live with it. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: gnu tar-1.15+ for RHEL4
On Fri, Apr 07, 2006 at 08:21:15PM -0400, Gene Heskett enlightened us: 2006/4/7, Gene Heskett [EMAIL PROTECTED]: I'm told that the 1.14 versions from RH, if fresh, have the backported patches that fixed 1.14 so it worked. I think I'd make a softlink see if it works. Yes the 1.14 fromRH is supposed to work, but I don't see how a symlink would resolve the issue ? Or I do not understand what you mean by a symlink... a symlink to the 1.14 package named tar-1.15 ? How would you create that and how would it fix the problem ??? I forgot to answer this one, see the manpage of ln, using the -s option in particular. That won't solve the RPM dependency. For that, he can rebuild the Fedora tar-1.15.1.src.rpm. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: snapshot volume using LVM2 and gnutar as dump program timeout?
On Thu, Apr 06, 2006 at 04:49:09PM +0200, Dennis Ortsen enlightened us: I'm using amanda 2.4.5 (RPM package from Fedora Core 4). I've got amanda running for a few months now, and am ready to move all hosts from the old backup server (BackupExec) to the server running amanda. On a few servers we used to work with LVM snapshot volumes to backup our data. When we configure amanda to backup a snapshot volume using gnutar, we seem to have a problem with the planner timing out on the estimate. The planner needs the estimate in order to make a solid decision on which disk to backup when and what level of backup is needed. All the other backups run fine, but whenever I want to backup a snapshot volume created with LVM2, I get a long timeout which eventually times out... Someone any idea's? Try running the commands by hand on a snapshot and see what takes so long. You'll find the commands being run in the logs, sendsize and sendbackup specifically. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: user $PATH problem
On Wed, Apr 05, 2006 at 10:20:59AM -0400, Gene Heskett enlightened us: When doing it as amanda, with amanda's full $PATH, /usr/local/sbin, where all of amanda's executables live, is NOT in the $PATH. Adding it to ~/.bash_profile seems to allow it to survive the pathmunge'ing being done in /etc/profile, so I'm A) confused as to why it does, and B) in any event, is there a good reason to dis-allow access to /usr/local/sbin for the normal user? Explain it to me please. Generally speaking, {foo}/sbin directories are system commands that normal users shouldn't/needn't be running. There are obvious exceptions to the rule, but you'll find at least on RedHat-like systems that no normal user has access to /sbin, /usr/sbin, etc. by default. For those special cases, you can just add it to the path of the necessary user as you did with the .bash_profile. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: how to split partitions
On Fri, Mar 31, 2006 at 01:26:31PM +0200, listrcv enlightened us: Stefan G. Weichinger wrote: As promised, I updated the wiki docs on this topic, see http://wiki.zmanda.com/index.php/Backing_up_large_amounts_of_data. Thanks for contributing, I just added some minor corrections. Thanks, that looks better now :) One thing I think that is missing is an explanation of the difference between: include include append exclude exclude append I would guess that 'append' tells amanda something like 'you should append this to the previous dumps of that mountpoint rather than considering it seperate', but I don't know. Also, the self-consistency is dubious since amanda seems to consider /home/wanda as a part of /home --- or maybe not. It would be nice if someone could shed a light on the logic behind that. It merely means append this to the previous {in,ex}clude I just gave you. Makes for easier reading. If you did: include ./foo include ./bar your include list would be ./bar (second include overrides the first) If you did: include ./foo include append ./bar your include list would be ./foo ./bar (second include is appended to the first) Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: amanda 2.5.0 : tar 1.15 required ?
On Fri, Mar 31, 2006 at 10:14:54AM -0800, Paddy Sreenivasan enlightened us: tar 1.15.1 is known to work well when Amanda. There are some versions of GNU tar that does not work well. See GNU tar note in http://wiki.zmanda.com/index.php/Installation#Dependencies That doesn't mean that 1.15 should be required by the spec file. tar 1.14 from Redhat works just fine (see https://www.redhat.com/advice/speaks_backport.html) On 3/31/06, FM [EMAIL PROTECTED] wrote: Hello, I try to rebuild amanda 2.5 with srpm from : http://www.zmanda.com for RedHat Enterprise 3 the spec ask for tar 1.15. I can easily change that but before that I just what to be just that it is a requirement reg -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: amanda 2.5.0 : tar 1.15 required ?
On Fri, Mar 31, 2006 at 11:52:04AM -0700, Michael Loftis enlightened us: --On March 31, 2006 1:36:36 PM -0500 Matt Hyclak [EMAIL PROTECTED] wrote: That doesn't mean that 1.15 should be required by the spec file. tar 1.14 from Redhat works just fine (see https://www.redhat.com/advice/speaks_backport.html) I don't know about Redhat tar 1.14 but GNU tar 1.14 is severely broken. It 'seems' to create ok archives, but in reality a lot of the time creates garbage even it can't read. Tan into this quite a bit, one symptom is the 'invalid base64 header' error. Right, Redhat has backported the fixes from 1.15.1 to 1.14. Which is why using a version number to ensure compatability is not a good idea IMHO. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: Extracting files error
On Thu, Mar 30, 2006 at 12:20:59PM -0600, Gordon J. Mills III enlightened us: When trying to extract files using amanda I get the following error. I have substituted hostname for the actual hostname. Any hints/ideas on what the problem is? Amanda ver: 2.4.5p1 Tape drive is 35G autoloader Extracting files using tape drive changer on host cuu-router.cuu.crownmx.com. The following tapes are needed: TAPE03 amrecover list TAPE TAPE03 LEVEL 2 DATE 2006-03-25 /data/IT/VirtualStats/VirtualStats.cfg amrecover extract Extracting files using tape drive changer on host hostname The following tapes are needed: TAPE03 Restoring files into directory /root Continue [?/Y/n]? Y Extracting files using tape drive changer on host hostname Load tape TAPE03 now Continue [?/Y/n/s/t]? Y ./data/IT/VirtualStats/VirtualStats.cfg tar: ./var/log/lastlog: invalid sparse archive member tar: Skipping to next header tar: Error exit delayed from previous errors extract_list - child returned non-zero status: 2 Continue [?/Y/n/r]? What version of tar are you using? Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: Install of 2.5.0 fails somehow
On Thu, Mar 23, 2006 at 11:17:26PM -0500, Gene Heskett enlightened us: Greetings; Deciding to bite the bullet and see if 2.5.0 was backwards compatible, I built it with the same config options I use now for a couple of years, building it as the user amanda, but installing it as root. Running an 'amcheck Daily' as amanda gets me this: [EMAIL PROTECTED] amanda-2.5.0-20060323]$ amcheck Daily amcheck: error while loading shared libraries: libamserver-2.5.0-20060323.so: cannot open shared object file: No such file or directory However: [EMAIL PROTECTED] amanda-2.5.0-20060323]$ ls -l /usr/local/lib|grep server -rwxr-xr-x 1 root root 399445 Dec 19 21:55 libamserver-2.4.5p1-20051218.so -rwxr-xr-x 1 root root 452388 Mar 23 23:06 libamserver-2.5.0-20060323.so -rw-r--r-- 1 root root 836784 Mar 23 23:06 libamserver.a -rwxr-xr-x 1 root root 884 Mar 23 23:06 libamserver.la lrwxrwxrwx 1 root root 29 Mar 23 23:06 libamserver.so - libamserver-2.5.0-20060323.so Is there anything I can check before the run starts, or should I just re-install the older one? Did you re-run ldconfig? Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: The server on the other side of my firewall is timming out agian.
On Fri, Mar 17, 2006 at 10:54:06AM +, Chuck Amadi Systems Administrator enlightened us: FAILURE AND STRANGE DUMP SUMMARY: server / lev 0 FAILED [server.my.co.ukNAK : amandad busy] tapeserver.sm /var lev 0 STRANGE It looks to me like an amanda process is already running/hung on server.my.co.uk. Can you verify that with ps? Thus this was working fine but for some unknown reason the timeout has exceeded 300 seconds Here is my log using less command /tmp/amanda/amandad.20060316205008000.debug file amandad: time 600.049: received other packet, NAKing it addr: peer 193.131.77.174 dup 193.131.77.174, port: peer 63313 dup 63366 amandad: time 600.182: sending nack: Amanda 2.4 NAK HANDLE 003-60610808 SEQ 1142538305 ERROR amandad busy Negative acknowledgement. amandad must already be running. amandad: time 617.155: sending REP packet: Amanda 2.4 REP HANDLE 003-60610808 SEQ 1142538305 OPTIONS features=feff9ffe0f; / 0 SIZE 5303320 / 1 SIZE 1568330 amandad: time 627.154: dgram_recv: timeout after 10 seconds amandad: time 627.154: waiting for ack: timeout, retrying amandad: time 637.154: dgram_recv: timeout after 10 seconds amandad: time 637.154: waiting for ack: timeout, retrying amandad: time 647.154: dgram_recv: timeout after 10 seconds amandad: time 647.154: waiting for ack: timeout, retrying amandad: time 657.154: dgram_recv: timeout after 10 seconds amandad: time 660.933: waiting for ack: timeout, retrying amandad: time 670.924: dgram_recv: timeout after 10 seconds amandad: time 670.924: waiting for ack: timeout, giving up! amandad: time 670.936: pid 13540 finish time Thu Mar 16 21:01:19 2006 Thus I assume that I need to amend my IPCHAIN rule from ipchains -M -S 7200 10 300 to ipchains -M -S 7200 10 800 I think this is a red herring, but I could be wrong. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: how to split partitions
On Fri, Mar 17, 2006 at 01:36:58PM +0100, listrcv enlightened us: Paul Bijnens wrote: Try this: prometheus /home/ALL_a /home { tar-dfault-no-x include ./a* } 1 prometheus /home/ALL_b /home { tar-dfault-no-x include ./b* } 1 ... etc etc ... prometheus /home/ALL_z /home { tar-dfault-no-x include ./z* } 1 prometheus /home/REST/home { tar-dfault-no-x exclude append ./[a-z]* } 1 The following does _not_ work as it should: prometheus /share_dbvnr /share { tar-dfault include ./db-vnr } 1 prometheus /share_edv /share { tar-dfault include ./edv } 1 prometheus /share_programs /share { tar-dfault include ./programs } 1 prometheus /share_rest /share { tar-dfault exclude append ./cpc ./cs ./data ./db-vnr ./edv ./programs } 1 It will dump the subdirectories as specified, but excluding the subdirectories results in another dump of everything under /share which I don't want because it will no longer fit on a single tape soon. This is probably because tar is being told to archive /share and to exclude the directories listed which will result in archieving them nonetheless, exactly as it was in the testing of tar options I did yesterday. How do I catch the rest? It does work for backing up /home: [...] prometheus /home_y /home { tar-dfault include ./y* } 1 prometheus /home_z /home { tar-dfault include ./z* } 1 prometheus /home_rest /home { tar-dfault exclude append ./[a-z]* } 1 How comes that it works for /home but not for /share? I have stopped using wildcards on /share because I'm getting dumper: FATAL error [dumper PORT-DUMP: too many args: 19 != 12] in the report. My idea was that spaces in filenames give me trouble, but not using wildcards didn't help neither the problem with the number of arguments, nor dumping /share twice. How can I find out what causes the problem with the arguments? GH I'd look at the sendsize.* files on the server. The .debug will show you the tar commands being run, the .include and .exclude files will show you the contents of the includes/excludes. See if you can track down the difference. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: how to split partitions
On Fri, Mar 17, 2006 at 01:56:16PM +0100, listrcv enlightened us: Matt Hyclak wrote: I'd look at the sendsize.* files on the server. The .debug will show you the tar commands being run, the .include and .exclude files will show you the contents of the includes/excludes. See if you can track down the difference. Good to know, thx! :) Just after I sent the mail, I got the idea to use seperate 'exclude append' lines for each directory to be excluded: prometheus /share_data_rest /share/data { tar-dfault exclude append ./archiv exclude append ./chaotisch exclude append ./entwicklung exclude append ./fibu exclude append ./materialwirtschaft exclude append ./qualitaetswesen exclude append ./Transfer exclude append ./vertrieb } 1 This looks much better in amstatus and seems to do what it should. If it works, I'll add a notice about using 'exclude append' in the wiki (if I can edit there). Interesting. I don't have any DLE's with more than one exclude append, so I never ran into that. Good to know, though! Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: tapetype w/compression for a HP Ultrium LTO 1 drive
On Fri, Mar 17, 2006 at 12:00:04PM -0500, Salada, Duncan S (Titan) @ TITAN enlightened us: I'm having trouble getting amtapetype to give me a 200gb tapetype for a HP Ultrium LTO 1 drive. It tells me that hardware compression is on, and I've tried using -e200g with -f/dev/rmt/0ubn and -f/dev/rmt/0cbn in desperation. But it still will only give me the 100gb tapetype: define tapetype Ultrium_I-HP_C7145_233S { comment just produced by tapetype prog (hardware compression on) length 100864 mbytes filemark 0 kbytes speed 14228 kps } I'm using amanda-2.4.5p1 on Solaris 10. I've searched around for a while but I've found nothing. Any hints/explanations would be greatly appreciated. An LTO 1 tape will only hold 100GB of data, period. 200GB is marketing hype for what they think you can get pre-compression. If you are going to use hardware compression, you'll just have to guess how compressible your data is and adjust the length accordingly. Might take some experimenting to find the right amount. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: Sizes don't fit?
On Thu, Mar 16, 2006 at 02:52:06PM +0100, listrcv enlightened us: Paul Bijnens wrote: On 2006-03-16 13:48, listrcv wrote: prometheus:/tmp/xxx# tar -cf test.tar -T include . When using -T you should not specify a list of things to backup on the command line as well, i.e. omit the final dot. Oh! :) Maybe I just have to try it out ... :/ Yes, indeed! I'm somewhat stuck with it. I need some doc on the format of the disklist to understand the example in the FAQ: fatboy /bigmama_BIGDIR /bigmama { # a big subdirectory comp-user-tar include ./bigdir } fatboy /bigmama_FILES01 /bigmama { # all files beginning with... nocomp-user-tar include ./file[01]* } The 'fatboy' is the server? What is bigmama_BIGDIR and bigmama_FILES01? And why is every thing called /bigmama? /bigmama is the mount point on the server. /bigmama_FILES01 is just a name given to differentiate those specific files on /bigmama from those included in /bigmama_BIGDIR Using prometheus /home/a*tar-dfault-no-x prometheus /home/b*tar-dfault-no-x [...] doesn't work, amcheck complains: ERROR: prometheus: [could not access /home/z* (/home/z*): No such file or directory] ERROR: prometheus: [could not access /home/y* (/home/y*): No such file or directory] [...] Right, those files don't exist. (Literal *) Would it be something like this? prometheus /home_a /home { tar-dfault-no-x include /home/a* exclude /home/[!a]* } prometheus /home_b /home { tar-dfault-no-x include /home/b* exclude /home/[!b]* } [...] prometheus /home_rest /home { tar-dfault-no-x include /home/[!a-z]* exclude /home/[a-z]* } Close. You don't need to exclude unless there something in your range you don't want. Here's what I do: my.fqdn.com /export_AI /export { # Directories A-I comp-user-tar include ./[a-i]* } my.fqdn.com /export_JQ /export { # Directories J-Q comp-user-tar include ./[j-q]* exclude ./profiles } my.fqdn.com /export_RZ /export { # Directories R-Z comp-user-tar include ./[r-z]* } my.fqdn.com /export /export { # All other directories - catchall comp-user-tar exclude ./[a-z]* exclude append ./profiles } Note /export_JQ where I want to ignore the profiles directory, but profiles is part of j-q. Same goes for the catchall at the end, I backup profiles separately so I want to exclude it as well as all of the already known quantities. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: Installation Configuration: disk definitions general questions
On Thu, Mar 16, 2006 at 04:41:04PM +, Peter Farrell enlightened us: Thanks you gentlemen all - for your comments and advice. The advice to RTFM is understood and well taken. The clarification as to the prepending of my exclude list especially... My hosts are set up and the backups ran smoothly - all reports minus AMANDA MAIL REPORT FOR BogusMonth 0, 0 ran fine. The 'bogus-month' one doesn't seem to be able to see it's log file... Anyway. Went away to my client-hosts and was trying to run: # amrecover misc_backups and have errors from both hosts... (but it runs as one would expect it to from the master-host (VEGA) - [EMAIL PROTECTED] tmp]# /usr/local/amanda/sbin/amrecover misc_backups AMRECOVER Version 2.4.5p1. Contacting server on vega.scarceskills.com ... 220 vega AMANDA index server (2.4.5p1) ready. 200 Access OK Setting restore date to today (2006-03-16) 200 Working date set to 2006-03-16. Scanning /home/data/amanda/dumps... 200 Config set to misc_backups. 200 Dump host set to vega.scarceskills.com. Trying disk /tmp ... $CWD '/tmp' is on disk '/tmp' mounted at '/tmp'. 200 Disk set to /tmp. /tmp amrecover - [EMAIL PROTECTED] etc]# /usr/local/amanda/sbin/amrecover misc_backups AMRECOVER Version 2.4.5p1. Contacting server on sabik.scarceskills.com ... amrecover: cannot connect to sabik.scarceskills.com: Connection refused -- [EMAIL PROTECTED] /]# /usr/local/amanda/sbin/amrecover misc_backups AMRECOVER Version 2.4.5p1. Contacting server on SSPC20.scarceskills.com ... amrecover: cannot connect to SSPC20.scarceskills.com: Success The clients are trying to contact themselves for the index service. I suspect that means you are using distro-provided RPMs and they have localhost compiled as the index server. Try amrecover -s your.tape.server misc_backups If you find it annoying, rebuild amanda on the clients. If you're using an RPM-based distro, see http://www.math.ohiou.edu/~hyclak/casit/amanda/ for some hints. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: Installation Configuration: disk definitions general questions
On Thu, Mar 16, 2006 at 11:50:26AM -0500, Matt Hyclak enlightened us: The clients are trying to contact themselves for the index service. I suspect that means you are using distro-provided RPMs and they have localhost compiled as the index server. I meant distro-provided packages. No offense to any Debianites :-) -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: New AMANDA server cannot one reach old client. (hostname lookup failed with the wrong IP)
On Mon, Mar 13, 2006 at 03:46:13PM -0500, Byarlay, Wayne A. enlightened us: I have a RHE4 machine with the standard AMANDA config, and then I copied over my amanda.conf file, disklist, and chg-conf file from the old AMANDA server. Now, all the clients report OK, except one! Stupid old AIX! Amcheck says, ERROR: stupid.server.edu [addr 555.555.555.555: hostname lookup failed] Obviously I replace the true name IP with fake names. The IP is NOT CORRECT! Even in the /etc/hosts file, it has the correct IP for this stupid.server. Any ideas on why amanda would think that stupid.server is the wrong IP? This is usually a reverse-DNS issue. What does host stupid.server.edu return? Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: New AMANDA server cannot one reach old client. (hostname lookup failed with the wrong IP)
On Mon, Mar 13, 2006 at 04:44:51PM -0500, Byarlay, Wayne A. enlightened us: Thanks all for your help, Did I mention that I'm somewhat newbie-ish with *nix? Blush. Thanks, here are my responses... 1. I did not know about this host command. It yields the INCORRECT information! snip -- snip -- snip 3. Stupid.aix.client can properly nslookup smart.amanda.server. Do I need to switch that nsswitch.conf thing? No, you need to fix DNS so that all the IPs map to the right hostnames and vice-versa. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: If most recent backup is not level 0, recovery fails to bring back all files when directories have been renamed
On Wed, Mar 08, 2006 at 10:02:43AM -0500, Gene Heskett enlightened us: On Wednesday, 08.03.2006 at 07:35 -0500, Gene Heskett wrote: On the client being backed-up: $ tar --version tar (GNU tar) 1.14 We believe this version of tar is borked. Please get the older 1.13-19, 1.13-25, or the newer 1.15-1. For whatever reason, 1.14 was only visible on gnu.org for about 5 or 6 weeks, being replaced with 1.15-1, which for gnu.org, considering the speed they normally run at, is instantainious. I've been running 1.15-1 since about a week after it became available without problems. Interesting. Do you think that the behaviour I'm seeing is solely as a result of 'tar' here? Is the behaviour I describe something that you believe *should* *not* happen with AMANDA when using 'tar'? I do not know this for a fact. ISTR the file headers it made weren't correct and recovery failures were the result. Possibly someone else can elaborate here? It created proper archives, however it failed extracting archives that had sparse files in them properly. RedHat still uses this version in RHEL4, but has backported the fix for this problem from 1.15.1. This package of 'tar' is the current default in Debian/Sarge and is updated with various security issues without changing the version number, so it's not entirely clear what version is really there under the hood. Which is exactly why I get this stuff (both tar and amanda) from the src and build my own. It Just Works(TM). :) I hope you build your own RPMs, since I know you use an RPM based system ;-) Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: If most recent backup is not level 0, recovery fails to bring back all files when directories have been renamed
On Wed, Mar 08, 2006 at 10:31:44AM -0500, Gene Heskett enlightened us: I hope you build your own RPMs, since I know you use an RPM based system ;-) Nope, I use checkinstall for some stuff, but this isn't being done for these. And they are pinned in yum.conf. Yeah, my systems borked. Funny thing is, it works fine for everything I want to do, which at times can best be described as 'an eclectic bunch' of apps. I know we're drifting a bit off-topic here, but usually what I do for non-mission-critical machines (such as a home workstation) is grab stuff from the latest fedora development repository and rebuild them on my system. That's what I did for tar, and what I do for amanda (even on production systems I do that). Just a suggestion :-) Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: controlling when a box is backed up.
On Tue, Mar 07, 2006 at 01:09:55PM -0700, Cameron Matheson enlightened us: I haven't ever learned too much about how amanda's scheduler works (mostly because it does just work), but now I have a situation which requires a bit more control. I have a box that needs to be backed up any time after 2:00am, but my backups start at 12:30am. Ideally I would like to keep my backups starting at 12:30. Is there a way to control what time a certain box is backed up? I checked the docs, but didn't see anything. Check the starttime parameter in amanda.conf. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: Preparing to migrate
On Sat, Mar 04, 2006 at 09:34:45AM -0500, Byarlay, Wayne A. enlightened us: folks, I'm getting to migrate my AMANDA from one server to another, and I was just wondering if there were any pitfalls I should be aware of. Let me describe my config: Current setup: AMANDA 2.4.4p1 on a RH9, backing up to Disk using the tape changer method. Installed using the downloaded amanda package; not an RPM or default RedHat install. Desired setup: RHE4, Amanda installed using the RHE4 first-time install, and I want to be backing up the same systems using the same config files, if possible. (amanda.conf location may have changed, etc.) The new AMANDA Host will have the SAME NAME AND IP ADDRESS as the old one. Any gotchas I should be aware of, client-wise or system-wise? I wouldn't recommend using the packaged RPM with RHEL4. At the very least, rebuild the SRPM to include proper references to your index and tape servers (the default is localhost, which is not a good thing). Alternatively, build the Fedora Development RPM on your RHEL 4 box to have the latest stable amanda. This is what I do and have been using in production for quite a while now. See http://www.math.ohiou.edu/~hyclak/casit/amanda/ for some hints. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: ip_conntrack_amanda problem with Linux kernel 2.6
On Fri, Mar 03, 2006 at 01:04:11PM +0100, Jorge Izquierdo (UAM) enlightened us: We are using amanda software to backup our servers and workstations onour department and we have a problem with the iptables configurations ofsome of the amanda clients. The problem is with the stations with Linux with kernel version 2.6. Using the same configuration as in Linux with kernel 2.4 for the iptables software the ones with kernel 2.6 reports an error when trying to make the backup because the server cannot connect to TCP ports suggested by the client. Those ports are not opened by default on the iptables configuration, the ip_conntrack_amanda module loaded from the /etc/sysconfig/iptables-config file, should open those ports (ramdomly chosed by the client) related to the first connection. So it seems that the ip_conntrack_amanda module on kernel 2.6 does not work properly. Any ideas? Any bug? One posible solution could be to open the range of ports from which client randomly select the port to dump the backup to server. Does anybody knows what this range is in the amanda-client RPM package or how to discover it? Thanks in advance if anybody may help me. What version of 2.6? There was an off-by-one type of error in Fedora Core 2's kernels at one point around 2.6.5, I think. ip_conntrack_amanda works just fine on my CentOS (RHEL) 4 machines, which is 2.6.9. All I do on clients is allow udp 10080 through and load ip_conntrack_amanda. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: ip_conntrack_amanda problem with Linux kernel 2.6
On Fri, Mar 03, 2006 at 02:39:50PM +0100, Jorge Izquierdo (UAM) enlightened us: We use Scientific Linux 4.2 with kernel 2.6.9-22.0.1.EL and also with Scientific Linux 4.0 2.6.9-5.0.5.EL the same problem Not sure what to tell you then, these kernels all worked on my CentOS boxes. You are allowing RELATED traffic back in to the box, right? Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: Intelegence of amflush
On Thu, Feb 23, 2006 at 02:31:06PM +0100, Paul Bijnens enlightened us: Paul Bijnens wrote: When there is only one image to choose from, Amanda will take that one, even if it will probably not fit. Changing that will make many other users of Amanda unhappy, e.g. those using hardware compression because their tape length is just a wild guess of the truth, or even me, because I underestimate the tape length a bit, so that amanda has a few percent margin with estimates being smaller than real dumps (that's why I get tapes filled with 106%). That's perfectly o.k. . But im my case there where about 25 images to choose from. But it will choose the correct one when you specify taperalgo largestfit. (The default is the first one found.) At least with 2.4.5, largestfit means of the oldest date, the largest that will fit. If you have 20GB tapes and a dump on Feb 15th that is 23GB and a dump on Feb 16th that is 13GB, it will pick the Feb 15th because it is older. Has this been changed in 2.5? Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: sendsize failures on Fedora Core 4
On Mon, Feb 20, 2006 at 05:17:16PM -0500, Drew Derbyshire enlightened us: I've run amanda for years on various FreeBSD machines. I THINK in days of old I also backed up at least one Redhat 8 client system, but the system retired a while back so I have no data and little memory. I'm now trying to add a Fedora 4 client, with poor results. amcheck runs fine, but the infamous estimate failed keeps happening when I actually try to run. It's not a timeout (total execution is only 30 seconds), and sendsize seems to be actively failing. I'm not sure what to make of it. In particular: * /sbin/dump 0Ssf 1048576 - /dev/sda1 works fine as the amanda user run from the command line. * GNU dump output hasn't changed much in years * /etc/dumpdates is world readable and group writable. Suggestions? sendsize: debug 1 pid 22313 ruid 33 euid 33: start at Mon Feb 20 13:17:20 2006 sendsize: version 2.4.5 sendsize[22315]: time 0.012: calculating for amname '/dev/sda1', dirname '/boot', spindle 1 sendsize[22315]: time 0.012: getting size via dump for /dev/sda1 level 0 sendsize[22315]: time 0.015: calculating for device '/dev/sda1' with 'ext3' sendsize[22315]: time 0.016: running /sbin/dump 0Ssf 1048576 - /dev/sda1 sendsize[22315]: time 0.029: running /usr/lib/amanda/killpgrp sendsize[22313]: time 0.037: waiting for any estimate child: 1 running sendsize[22315]: time 0.045: DUMP: cannot read /etc/dumpdates: Permission denied sendsize[22315]: time 0.047: DUMP: The ENTIRE dump is aborted. sendsize[22315]: time 0.049: . sendsize[22315]: estimate time for /dev/sda1 level 0: 0.032 sendsize[22315]: no size line match in /sbin/dump output for /dev/sda1 sendsize[22315]: . sendsize[22315]: estimate size for /dev/sda1 level 0: -1 KB sendsize[22315]: time 0.049: asking killpgrp to terminate sendsize[22315]: time 1.053: done with amname '/dev/sda1', dirname '/boot', spindle 1 sendsize[22313]: time 1.054: child 22315 terminated normally sendsize[22318]: time 1.055: calculating for amname '/dev/sda5', dirname '/', spindle 1 sendsize[22318]: time 1.055: getting size via dump for /dev/sda5 level 0 sendsize[22318]: time 1.057: calculating for device '/dev/sda5' with 'ext3' sendsize[22318]: time 1.057: running /sbin/dump 0Ssf 1048576 - /dev/sda5 sendsize[22318]: time 1.065: running /usr/lib/amanda/killpgrp sendsize[22313]: time 1.073: waiting for any estimate child: 1 running sendsize[22318]: time 1.087: DUMP: cannot read /etc/dumpdates: Permission denied If I had to guess, I'd say you've got SELinux enabled and it's interfering. Anything in /var/log/messages to tell if that's the case? Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: Migrating to Amanda, question 2
On Fri, Feb 10, 2006 at 01:40:38PM -0500, Joshua Baker-LePain enlightened us: On Friday 10 February 2006 11:42, Salvatore Enrico Indiogine wrote: tar: CentSO tar-1.14-8.RHEL4 FC4 tar-1.15.1-11.FC4 CentOS tar-1.14.* is known bad, I'm using 1.15-1 myself, with 1.13-25 installed as a fallback available with a rebuild/reinstall of amanda, a 6 minute job typically... Actually, centos-4's tar seems to work just fine. They recently fixed a sparse files bug in it, and it seems quite happy. I've recently done a few multi-TB backup/restores with it with no problem. Or is there something I'm missing? As to centos vs. FC, I prefer centos if only for the longer life cycle and the feeling that it's a bit more tested than FC. That being said, I've never run FC, so take that all with a grain of salt. I'll second Joshua here and point you to http://www.math.ohiou.edu/~hyclak/casit/amanda/ I recommend getting the lastest 2.4.5pX source RPM from Fedora and rebuilding it on your CentOS machine. That way you can specify tape servers and default configurations with options to rpmbuild. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: LTO2 Drive / Linux anyone ?
On Fri, Feb 10, 2006 at 03:47:51PM -0500, Guy Dallaire enlightened us: We're curently using a 5 slot DLT IV tape changer. I'm getting an increased amount of errors and the tape drive is quite old now. We're considering the purchase of s single LTO2 drive. We're using amanda 2.4.5 on centos 4.2 without problem Kernel is 2.6.9-22.0.1.ELsmp Gnu Tar is 1.14 I would like to know if any of you is using a similar OS for the amanda server, having succes with LTO2 drives and the drive manufacturer/model you are using. I'm using RHEL3 at the moment with 2.4.5p1 and a Centrance (now Quantum) 1U rackmount LTO2 drive. It's got room for a second drive in it and only cost about $2200. Way better than the $3500 HP wanted for a refurbed, 90-day warranty to replace the LTO1 drive that died a month out of warranty. The speed is a little slow on the 1U, but they have a 2U that is faster. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: LTO2 Drive / Linux anyone ?
On Fri, Feb 10, 2006 at 04:30:08PM -0500, Gene Heskett enlightened us: We're curently using a 5 slot DLT IV tape changer. I'm getting an increased amount of errors and the tape drive is quite old now. We're considering the purchase of s single LTO2 drive. We're using amanda 2.4.5 on centos 4.2 without problem Kernel is 2.6.9-22.0.1.ELsmp Gnu Tar is 1.14 That 1.14-* is a zinger for most folks, use 1.15-1 or 1.13-19 or 1.13-25 for best results. The 1.14 included in RHEL4 should be ok. See http://www.redhat.com/advice/speaks_backport.html When dealing with RHEL (or CentOS, WB, Tao, etc.), version numbers are not enough to determine problems. That being said, rebuilding the 1.15.1 srpm from Fedora works fine as well. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: data timeout error
On Thu, Feb 09, 2006 at 02:12:55PM -0500, Guy Dallaire enlightened us: One of my DLE fails to backup correctly from time to time. It happened twice this week. This is a bit annoying because it changes everyday (it's an oracle cold/hot backup destination for dabase files) and I really need to get this stuff backed up every day. This is the only DLE exhibiting this behavior. I have to use anoither tape in the morning and RE-BACKUP this DLE (Which normally works fine) It's not really huge. The error I have in the amanda report is: FAILURE AND STRANGE DUMP SUMMARY: sol /disk1/RDBMS_BACKUP lev 0 FAILED [data timeout] It's no really huge. I'n using amanda 2.4.5 the tape server is centos 4.2 (RHEL 4 Clone) and the client with the failing DLE is a sun solaris 9 box. Is there a timeout I can configure for this ? What might cause this ? The client is in the DMZ, but I have FW1 rules allowing for the backup UDP and TCP ports. Check the setting of your dtimeout value in amanda.conf. You might also want to check out http://www.amanda.org/docs/faq.html#id2551493 if you have a firewall configured on your CentOS box. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: data timeout error
On Thu, Feb 09, 2006 at 02:31:11PM -0500, Guy Dallaire enlightened us: 2006/2/9, Matt Hyclak [EMAIL PROTECTED]: Check the setting of your dtimeout value in amanda.conf. You might also want to check out http://www.amanda.org/docs/faq.html#id2551493 if you have a firewall configured on your CentOS box. Hum... My dtimeout parameter has already been bumped from 20 minutes to 40 minutes (2400 seconds) Does this mean that there might be a delay of 40 minutes between EACH PACKET (Which is totally ridiculous) or does that mean that the backup of a client (or a DLE) cannot take more than 40 minutes from start to finish ? If so, where do I look for debugging ? I believe once the client starts performing the backup, no data is sent until it is completed. If any DLE takes longer than dtimeout * numberofDLEs, you'll get that problem. I would start on the client and look at the log files. You should be able to see how long it takes to complete the backup. If it's longer than your dtimeout allows for, you'll know where the problem is. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: Backing up Maildirs
On Wed, Jan 25, 2006 at 05:18:51PM -0700, Cameron Matheson enlightened us: I've recently gotten the appropriate holes cut in my firewall that allow me to backup my mail server. The problem is that Maildirs don't seem to want to be backed up... the backups fail, w/ hundreds of errors like these in the logs: ?gtar:./home/vmail/fjcomm.com/V/foo.bar/new/1137800711.V805Iac017.zena.tonservices.com: Warning: Cannot stat: No such file or directory I tried googling for what exactly 'Cannot stat' means, but I didn't come up w/ anything too conclusive. Anyway, why exactly should there be so many hundreds of these erros (mostly confined to one or two users mailboxes (but different users every night). If i don't back up the Maildirs, that server backs up just fine. It doesn't seem like these mailboxes should be changing *that* much overnight (maybe i'm wrong?). Maildir moves the files from the new/ directory to the cur/ directory when they've been accessed. Usually this means that the user (or the user's mail program) was running between the estimates being done and the backup being performed. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: IPv6 problem (on IPv4 system)?
On Wed, Jan 18, 2006 at 02:21:28PM -0500, Ken D'Ambrosio enlightened us: [Apologies if this is a duplicate; I sent it yesterday, and haven't seen it crop up yet...] Hi, all. Just recently, I added an Ubuntu Linux box to my series of Debian servers. All the stock Debian servers had backed up fine. However, my Ubuntu box fails miserably. Going through the client's debug log, the thing that really stands out is this: amandad: time 0.000: sending ack: Amanda 2.4 ACK HANDLE 001-B8CA0608 SEQ 1137474902 amandad: time 0.000: dgram_send_addr: sendto(0.0.0.0.858) failed: Invalid argument A strace run on the client's amandad shows the following interesting snippet: sendto(0, Amanda 2.4 REP HANDLE 001-F0B70608 SEQ 1137547581\nERROR [addr 0.0.0.0: hostname lookup failed]\n, 95, 0, {sa_family=AF_INET6, sin6_port=htons(847), inet_pton(AF_INET6, ::58d0:f6bf:c927:f6b7, sin6_addr), sin6_flowinfo=0, sin6_scope_id=134513844}, 16) = -1 EINVAL (Invalid argument) It looks to me like the client -- which *is* IPv6 enabled, but has no IPv6 interfaces -- is thinking IPv6-ish thoughts. Some Googling showed a thread on this back in '03, and implied that a patch had been merged. Since I'm running 2.4.5, I would have thought that would be yesterday's news. Any ideas on what to do? Someone was commenting about this in my local LUG IRC channel that Ubuntu wasn't using xinetd, but something else called inetutils-inetd. He was getting the same error and apparently fixed it by installing xinetd. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: beep! (encryption, multiplexing...)
On Thu, Dec 29, 2005 at 04:18:09PM -0800, Kevin Till enlightened us: I agree with Paddy that ssh provides transport encryption and authentication. The only caveat is that the amanda binary needs to be installed at the same location in the server as well as in the client since server is running: /path/ssh -l CLIENT_LOGIN client.zmanda.com $libexecdir/amandad to start the backup process. That's a problem (at least on RedHat/Fedora) for 64-bit users, since $libexecdir is /usr/lib64 as opposed to /usr/lib on 32-bit systems. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: Volume is not in dump format
On Thu, Dec 22, 2005 at 01:37:23PM -0500, Myron Kowalski enlightened us: I've never had a problem restore files to solaris boxes; however, I'm trying to restore a linux backup, which I've never done. This is the message I got. amrestore: 79: restoring aimserver._dev_hda2.20051216.0 Volume is not in dump format amrestore: 80: reached end of information I see the tape has the image I want. I usually use this command: amrestore -p (drive) aimserver /dev/hda2 | ufsrestore ivf - Is the error telling me that linux has the compressed the image a different way. This is the line from disklist aimserver /dev/hda2linux-tar # /usr The image was backed up using tar, and you're trying to restore with ufsdump. That does not compute. Try: amrestore -p (drive) aimserver /dev/hda2 | tar xvf - or similar. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: /etc/dumpdates
On Tue, Dec 20, 2005 at 11:09:08AM -0500, Jay Fenlason enlightened us: On Mon, Dec 19, 2005 at 06:47:15PM -0500, Paul Seniuk enlightened us: Matt, Well you were right and that worked. Annoying story to it ...collegue decided to upgrade the box to FC4 and not tell me. The upgrade turned SELinux on by default. Merry Xmas Matt and thanks :) Sounds like there needs to be some work done on the selinux definitions. /me pokes Jay /me points Matt at Dan Walsh et all. :-) I'm running my Amanda server on a Rawhide box, and I don't see many AVC denied messages in my logs, so maybe they've improved the policy for FC5. I haven't tried amanda on a RHEL4 box with SELinux on yet or not. I'll try to get to that when I have a spare tuit, but will file a bug report if necessary. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: /etc/dumpdates
On Mon, Dec 19, 2005 at 05:45:45PM -0500, Paul Seniuk enlightened us: Perms on /etc/dumpdates is: -rw-rw-r-- 1 root disk 172 Dec 16 02:37 dumpdates Would anything be logged about failing to create /etc/dumpdates (get that long pole out, I used the RPM version for CentOS) ? For 'fun', I tried putting the perms to 777 ..still same error Any feedback on this would be appreciated :) Do you by chance have SELinux enabled on this machine and not on the others? Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: /etc/dumpdates
On Mon, Dec 19, 2005 at 06:47:15PM -0500, Paul Seniuk enlightened us: Matt, Well you were right and that worked. Annoying story to it ...collegue decided to upgrade the box to FC4 and not tell me. The upgrade turned SELinux on by default. Merry Xmas Matt and thanks :) Sounds like there needs to be some work done on the selinux definitions. /me pokes Jay Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: Which ports to open in which direction...
On Tue, Dec 06, 2005 at 10:01:37AM +0100, Paul Bijnens enlightened us: David Leangen wrote: http://wiki.zmanda.com/index.php/Configuration_with_iptables How does the ip_conntrack_amanda kernel module fits in here? I think that just using that module simplifies a lot of the setup. I'm not sure sure it handles amrecover connections though... I just ran amrecover from a client outside my firewall with ip_conntrack_amanda handling everything just fine. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: Why do I need a dumpuser account?
On Sat, Dec 03, 2005 at 10:18:49PM -0700, [EMAIL PROTECTED] enlightened us: I started reading the Setup the Backup Client Hosts section of the Amanda docs, which immediately assumes that I have a folder ~dumpuser/. It's telling me to create an .amandahosts file within that directory. Portage did not create any such user. Is this a user that I need in addition to the amanda user? The client and server are the same machine in my case (for now). I think dumpuser is just the generic term. In your case dumpuser == amanda. Other systems use backup or operator. Just replace ~dumpuser with ~amanda for your configuration. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: amrecover from the DMZ
On Fri, Dec 02, 2005 at 04:16:07PM -0800, Kevin Till enlightened us: Matt Hyclak wrote: On Thu, Nov 10, 2005 at 02:32:41PM -0500, Guy Dallaire enlightened us: 2005/11/10, Matt Hyclak [EMAIL PROTECTED]: Which ports does amrecover use to contact the tape server ? 10082 tcp and 10083 tcp. amrecover needs a privileged tcp port to connect to the server. It's the only amanda program that requires privileged TCP port I read that the other way around as what port *on* the server. The priviledged port on the client is what necessitates running amrecover as root. Glad we got that cleared up :-) Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: Edit tapelist to change tape order
On Thu, Dec 01, 2005 at 09:53:26AM +0100, Montagni, Giovanni enlightened us: I have a problem with tapelist file (shown below) 20051123 bkalfa-all07 reuse 20051123 bkalfa-all06 reuse 20051116 bkalfa-all04 reuse 20051110 bkalfa-all03 reuse 20051110 bkalfa-all02 reuse 20051109 bkalfa-all00 reuse 0 bkalfa-all05 reuse 0 bkalfa-all01 reuse snip -- snip -- snip This is not an ordered tapelist file (as i want, to make tape changing every night more simple). What's more simple than putting in the tapes amcheck lists in the e-mail to you? Amanda has no concept of order. I have tapecycle = 8 but i've also tried with a lower value (4 or less) and nothing has changed, amadmin claim that amanda want tapes labelled bkalfa-all01 - bkalfa-all05. That said, it looks like you could do the following: * Change your tapecycle to something low. 4 should be fine. * On your next run, bkalfa-all00 should be old enough to be used again. Put bkalfa-all00 and bkalfa-all01 in the drive. * On the next run, use bkalfa-all02 and bkalfa-all03 * On the next run, use bkalfa-all04 and bkalfa-all05 * Return your tapecycle to 8, continue life as normal Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: Next tape?
On Thu, Dec 01, 2005 at 06:53:31AM -0800, Yan Seiner enlightened us: How do I query amanda for the next tape? I run my backups on my home server; at times I am away from home so I can't feed the tapes in daily. I would like to be emailed a reminder telling me which tape is next. Right now amanda tells me which tape it expects when it completes a successful backup; I want to query it so it tells me on demand amcheck will tell you, or alternatively amadmin CONFIGNAME tape Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: Changing User when using RPMs
On Mon, Nov 28, 2005 at 05:35:16PM -0500, Gene Heskett enlightened us: Jeff: How do the rpms handle the perms? When building from tarballs, we always do the configuration building as the user amanda, and then become root to do the install so the perms are always handled correctly. Root is also required to run ldconfig so the libraries can be found after the install also. Does the rpm install do that? Currently, the amanda username is hardcoded, although I think Jay was in the process of changing that this afternoon in the Rawhide RPM to be configurable without modifying the spec file. The permissions are set at packaging time, not build time, since you should build rpms as non-root, but chances are you won't build as the amanda user. So when collecting all the files, there are a bunch of lines like: %attr(-,amanda,disk)somefile That changes the ownership to amanda:disk, and doesn't change the permissions. The RPM also creates the amanda user as well as running ldconfig after install. If you're versed in .spec files at all, take a peek at the one in rawhide. It's shaping up to be pretty usable. In fact, I think I'll stop maintaining my own since most of my local changes are in those RPMs now. I'll just rebuild from there when I need a new version. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: Changing User when using RPMs
On Mon, Nov 28, 2005 at 07:35:53PM -0700, Broderick Wood enlightened us: My resistance comes from the fact that we have been using AMANDA for about 2 years whereas we have had a user called amanda for about 15 years... I also don't want to do a source build as we are maintaining about 500+ machines in varying state of OS version/patch/etc... We use CfEngine to push out our localizations so I can fairly easily push a RPM to a machine and have been doing so which works fine except for TAR backups which is where I run into the hard-coded problem. So I am building a custom RPM and we'll see how that goes... There goes the easy maintenance of YUM and such for updating it... One of the suggestions from a *BSD person was to follow the same convention of naming system accounts _username instead of username, so that when Amanda Smith has an account on your system as amanda, you don't collide. I believe that's in a bugzilla on redhat's site, you might voice your opinion there if that is something you'd be interested in. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: Changing User when using RPMs
On Sun, Nov 27, 2005 at 08:48:09PM -0700, Broderick Wood enlightened us: Is there anyone out there who is using the RedHat RPMs abut has to change the user that runs the backups? ie. I want to change the name from amanda to backup or some other ID but runtar seems to have the id amanda hard coded into it. See http://www.math.ohiou.edu/~hyclak/casit/amanda - specifically the Recompiling Amanda via RPM. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: Backing up ACL's ?
On Fri, Nov 25, 2005 at 09:35:47AM -0500, Guy Dallaire enlightened us: The network admin here would like to start using ACL's on some files (Centos 4.1 a read hat 4 clone, and Red Hat 4). We use amanda (with gnutar) I don't think gnu tar supports it (other than by using dump, which we don't want to do for reliability reasons) Do you know if ACL support is planned ? Not directly. You have a couple options: 1. Dump the ACLs to a file (probably with getfattr) and back up that file. 2. Write a wrapper for gnutar that uses star, which does support some ACLs last I checked. There's probably others, as well, but it's late :-) HTH, Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: amrecover index empty
On Wed, Nov 23, 2005 at 02:48:14PM +0100, Jehan PROCACCIA enlightened us: I use amnda to backup several servers, unfortunatly since a few weeks ago, indexes for amrecover don't work anymore. I check the faq-o-matic http://amanda.sourceforge.net/fom-serve/cache/24.html and I am not in any cases mentioned index yes is set, amrecover is call with -C , and I use Gnu tar tar-1.14-4 on my fedora core 3 backup systeme ( kernel 2.6.10-1.770_FC3 amanda-server-2.4.4p3-4) tar 1.14 is not a good version of tar to use. See http://www.amanda.org/docs/faq.html#id2554919 Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263
Re: newbie Q: DUMP: The ENTIRE dump is aborted.
On Sat, Nov 19, 2005 at 09:21:52AM -0800, Yan Seiner enlightened us: I've set up amanda as best as I know how amcheck reports everything is OK. Permissions, etc. all seem OK. But when I try to run amdump I get: FAIL dumper localhost /home/mysql 20051119 0 [/sbin/dump returned 1] sendbackup: start [localhost:/home/mysql level 0] sendbackup: info BACKUP=/sbin/dump sendbackup: info RECOVER_CMD=/usr/bin/gzip -dc |/sbin/restore -f... - sendbackup: info COMPRESS_SUFFIX=.gz sendbackup: info end | DUMP: You can't update the dumpdates file when dumping a subdirectory | DUMP: The ENTIRE dump is aborted. sendbackup: error [/sbin/dump returned 1] INFO taper tape SeinerHome02 kb 0 fm 0 [OK] FINISH driver date 20051119 time 81.367 Any ideas on where to start? I don't want to clutter the list with long files So where do I need to look? Or what should I provide to the list so you can advise? Thanks, --Yan Well, the error is right there: sendbackup: info end | DUMP: You can't update the dumpdates file when dumping a subdirectory | DUMP: The ENTIRE dump is aborted. sendbackup: error [/sbin/dump returned 1] You are trying to use dump on a directory, not a mount point like it was designed. If you only want to do a subdirectory, you should use tar, not dump. Matt -- Matt Hyclak Department of Mathematics Department of Social Work Ohio University (740) 593-1263