Re: tape problem
amanda v 3.5.1, Debian AMD64 Buster, Quantum STO-5, LSI SAS2008 PCIe card I think I've found my problem. Amanda can't write big files to tape. >From my Debian kernel.log (grep for st0): Dec 13 15:41:34 sbox kernel: [ 2708.331836] st 0:0:2:0: [st0] Block limits 1 - 16777215 bytes. Dec 13 15:44:32 sbox kernel: [ 2885.931242] st 0:0:2:0: [st0] Error e (driver bt 0x0, host bt 0xe). Dec 13 15:44:32 sbox kernel: [ 2885.931245] st 0:0:2:0: [st0] Error on write filemark. Dec 13 15:44:38 sbox kernel: [ 2891.941684] st 0:0:3:0: Attached scsi tape st0 Dec 13 15:44:38 sbox kernel: [ 2891.941687] st 0:0:3:0: st0: try direct i/o: yes (alignment 4 B) Dec 13 15:46:28 sbox kernel: [ 3001.878038] st 0:0:3:0: [st0] Block limits 1 - 16777215 bytes. Dec 14 14:06:33 sbox kernel: [2.424988] scsi host0: ata_piix Dec 14 14:06:33 sbox kernel: [ 11.462017] st 3:0:0:0: Attached scsi tape st0 Dec 14 14:06:33 sbox kernel: [ 11.462018] st 3:0:0:0: st0: try direct i/o: yes (alignment 4 B) Dec 14 15:53:45 sbox kernel: [ 6603.276972] st 3:0:0:0: [st0] Block limits 1 - 16777215 bytes. Dec 14 16:08:29 sbox kernel: [ 7487.326410] st 3:0:0:0: [st0] Error 40 (driver bt 0inx0, host bt 0x0). Dec 14 16:09:05 sbox kernel: [ 7523.203323] st 3:0:1:0: Attached scsi tape st0 Dec 14 16:09:05 sbox kernel: [ 7523.203326] st 3:0:1:0: st0: try direct i/o: yes (alignment 4 B) and on and on for pages. dmseg says: 16777215 root@sbox:/var/log# dmesg | egrep st0 [2.424988] scsi host0: ata_piix [ 11.462017] st 3:0:0:0: Attached scsi tape st0 [ 11.462018] st 3:0:0:0: st0: try direct i/o: yes (alignment 4 B) [ 6603.276972] st 3:0:0:0: [st0] Block limits 1 - 16777215 bytes. [ 7487.326410] st 3:0:0:0: [st0] Error 40 (driver bt 0x0, host bt 0xin0). [ 7523.203323] st 3:0:1:0: Attached scsi tape st0 [ 7523.203326] st 3:0:1:0: st0: try direct i/o: yes (alignment 4 B) [ 7570.604960] st 3:0:1:0: [st0] Block limits 1 - 16777215 bytes. [ 7844.727097] st 3:0:2:0: Attached scsi tape st0 [ 7844.727101] st 3:0:2:0: st0: try direct i/o: yes (alignment 4 B) [ 7955.289508] st 3:0:2:0: [st0] Block limits 1 - 16777215 bytes. [130700.591001] st 3:0:2:0: [st0] Error e (driver bt 0x0, host bt 0xe). [130702.840754] st 3:0:2:0: [st0] Error 1 (driver bt 0x0, host bt 0x1).16777215 [130706.850766] st 3:0:3:0: Attached scsi tape st0 [130706.850787] st 3:0:3:0: st0: try direct i/o: yes (alignment 4 B) [131166.864674] st 3:0:3:0: [st0] Block limits 1 - 16777215 bytes. It looks to me like amanda's trying to write data that's too big for something in the kernel (there are 1G files in the holding disk). If that's correct, it looks like a simple change in some config should fix it. I've looked for a config to change, but can't find it. grep tells me that '16777215' (aka 0xff) isn't in any file in /boot. On the web, it looks like others are having similar troubles in other Linux distributions and with other backup software. That says to me that there was a change in my most recent update of the Linux kernel -- and that (probably) means the size definition is somewhere in the source (or maybe a binary number somewhere in /boot). I've compiled kernels before, so I think I could probably find this setting somewhere in the source, but the source is huge. Does anyone know where that 16777215 number is? Is it accessible in the make config file? Is it in one of the C include files? Do the amanda developers have a patch to deal with this? Is there a more civilized way to do this? Am I looking at the wrong thing? -- Glenn English
Re: tape problem
Am 06.12.21 um 15:39 schrieb David Simpson: Further to Jose's questions. Some config excerpt may help. On blocksize, that is defined within the tapetype. blocksize 512 kbytes And maybe try a run with amtapetype.
RE: tape problem
Further to Jose's questions. Some config excerpt may help. On blocksize, that is defined within the tapetype. blocksize 512 kbytes - David Simpson - Senior Systems Engineer ARCCA, Redwood Building, King Edward VII Avenue, Cardiff, CF10 3NB David Simpson - peiriannydd uwch systemau ARCCA, Adeilad Redwood, King Edward VII Avenue, Caerdydd, CF10 3NB -Original Message- From: owner-amanda-us...@amanda.org On Behalf Of Jose M Calhariz Sent: 06 December 2021 11:17 To: ghe2001 Cc: amanda-users@amanda.org Subject: Re: tape problem I think the explanation for most of your problems is explained the amanda report. I have some questions that may help find the problem: - how much free space do you have on holding disk, how much space is used? - How big are the backups every day? - Do you use hardware compression on the tape? Kind regards Jose M Calhariz On Fri, Dec 03, 2021 at 11:18:08PM +, ghe2001 wrote: > amanda version: amadmin-3.5.1 > OS: Debian Linux, Buster > Host: Supermicro > PCI card: LSI Logic / Symbios Logic SAS2008 PCI-Express Fusion-MPT > SAS-2 [Falcon] (rev 03) Tape drive: Quantum LTO-5 > Tapes: Quantum Ultrium 5 > > Problem: Over the years, I've written a bunch of amanda scripts. Here's what > the one that displays what's going on during a backup says at one point: > > gobook3.slsware.lan:/usr 20211203133758 0 4519450k > dump done, wait for writing > pi.slsware.lan:/lib20211203133758 0 1720720k > dump done, writing (1518944k done (88.27%)) (14:03:16) > pi.slsware.lan:/toshiba1 20211203133758 0 5089430k > dump done, wait for writing > pi.slsware.lan:/usr20211203133758 0 3108660k > dump done, wait for writing > sbox.slsware.lan:/blackHole/amanda 20211203133758 0 2251540k > dump done, wait for writing > sbox.slsware.lan:/home 20211203133758 0 74493810k > dump done, wait for writing > sbox.slsware.lan:/usr 20211203133758 0 5144210k > dump done, wait for writing >0: writing (pi.slsware.lan:/lib) === Fri Dec 03, 2021 > 02:03:35 PM > > This is normal. A lot of flushing's been done, and all the small files. > The script futzes with amstatus and displays every 5 seconds. > > . > > The '.' means amstatus has written the same line, and nothing of interest > has happened. > > gobook3.slsware.lan:/usr 20211203133758 0 4519450k > dump done, wait for writing > pi.slsware.lan:/toshiba1 20211203133758 0 5089430k > dump done, wait for writing > pi.slsware.lan:/usr20211203133758 0 3108660k > dump done, wait for writing > sbox.slsware.lan:/blackHole/amanda 20211203133758 0 2251540k > dump done, wait for writing > sbox.slsware.lan:/home 20211203133758 0 74493810k > dump done, wait for writing > sbox.slsware.lan:/usr 20211203133758 0 5144210k > dump done, wait for writing >0: tape error: Couldn't rewind device to finish: No > such device, splitting not enabled (pi.slsware.lan:/lib) === Fri Dec > 03, 2021 02:03:52 PM > > Notice the than ~30 seconds from the last display. And who's trying to > rewind and finding no device (/dev/nst0 I suppose). > > After this, every line says "terminated while waiting for writing." > > This looks to me like maybe something can't deal with files larger that 1G. > But I've seen it get to 55G or so. > > > This morning, Logwatch said: > > WARNING: Kernel Errors Present > st 0:0:3:0: [st0] Error 1 (driver bt ...: 2 Time(s) > st 0:0:3:0: [st0] Error e (driver bt ...: 1 Time(s) > st 0:0:4:0: [st0] Error 1 (driver bt ...: 1 Time(s) > st 0:0:4:0: [st0] Error e (driver bt ...: 1 Time(s) > st 0:0:5:0: [st0] Error 1 (driver bt ...: 1 Time(s) > st 0:0:5:0: [st0] Error e (driver bt ...: 1 Time(s) > st 0:0:6:0: [st0] Error 1 (driver bt ...: 1 Time(s) > st 0:0:6:0: [st0] Error e (driver bt ...: 1 Time(s) > st 0:0:7:0: [st0] Error 1 (driver bt ...: 1 Time(s) > st 0:0:7:0: [st0] Error 7 (driver bt ...: 1 Time(s) > st 0:0:7:0: [st0] Error e (driver bt ...: 1 Time(s) > > st0 is the non-rewinding tape drive. As best I know, nobody does anything > with st0 -- nst0 is what I, and amanda, use. > > This looks like there might be a bent driver. > > > Amcheck: > > slot 1: volume 'sls-9' > Will write to volume 'sls-9
Re: tape problem
I think the explanation for most of your problems is explained the amanda report. I have some questions that may help find the problem: - how much free space do you have on holding disk, how much space is used? - How big are the backups every day? - Do you use hardware compression on the tape? Kind regards Jose M Calhariz On Fri, Dec 03, 2021 at 11:18:08PM +, ghe2001 wrote: > amanda version: amadmin-3.5.1 > OS: Debian Linux, Buster > Host: Supermicro > PCI card: LSI Logic / Symbios Logic SAS2008 PCI-Express Fusion-MPT SAS-2 > [Falcon] (rev 03) > Tape drive: Quantum LTO-5 > Tapes: Quantum Ultrium 5 > > Problem: Over the years, I've written a bunch of amanda scripts. Here's what > the one that displays what's going on during a backup says at one point: > > gobook3.slsware.lan:/usr 20211203133758 0 4519450k > dump done, wait for writing > pi.slsware.lan:/lib20211203133758 0 1720720k > dump done, writing (1518944k done (88.27%)) (14:03:16) > pi.slsware.lan:/toshiba1 20211203133758 0 5089430k > dump done, wait for writing > pi.slsware.lan:/usr20211203133758 0 3108660k > dump done, wait for writing > sbox.slsware.lan:/blackHole/amanda 20211203133758 0 2251540k > dump done, wait for writing > sbox.slsware.lan:/home 20211203133758 0 74493810k > dump done, wait for writing > sbox.slsware.lan:/usr 20211203133758 0 5144210k > dump done, wait for writing >0: writing (pi.slsware.lan:/lib) > === Fri Dec 03, 2021 02:03:35 PM > > This is normal. A lot of flushing's been done, and all the small files. > The script futzes with amstatus and displays every 5 seconds. > > . > > The '.' means amstatus has written the same line, and nothing of interest > has happened. > > gobook3.slsware.lan:/usr 20211203133758 0 4519450k > dump done, wait for writing > pi.slsware.lan:/toshiba1 20211203133758 0 5089430k > dump done, wait for writing > pi.slsware.lan:/usr20211203133758 0 3108660k > dump done, wait for writing > sbox.slsware.lan:/blackHole/amanda 20211203133758 0 2251540k > dump done, wait for writing > sbox.slsware.lan:/home 20211203133758 0 74493810k > dump done, wait for writing > sbox.slsware.lan:/usr 20211203133758 0 5144210k > dump done, wait for writing >0: tape error: Couldn't rewind device to finish: No such > device, splitting not enabled (pi.slsware.lan:/lib) > === Fri Dec 03, 2021 02:03:52 PM > > Notice the than ~30 seconds from the last display. And who's trying to > rewind and finding no device (/dev/nst0 I suppose). > > After this, every line says "terminated while waiting for writing." > > This looks to me like maybe something can't deal with files larger that 1G. > But I've seen it get to 55G or so. > > > This morning, Logwatch said: > > WARNING: Kernel Errors Present > st 0:0:3:0: [st0] Error 1 (driver bt ...: 2 Time(s) > st 0:0:3:0: [st0] Error e (driver bt ...: 1 Time(s) > st 0:0:4:0: [st0] Error 1 (driver bt ...: 1 Time(s) > st 0:0:4:0: [st0] Error e (driver bt ...: 1 Time(s) > st 0:0:5:0: [st0] Error 1 (driver bt ...: 1 Time(s) > st 0:0:5:0: [st0] Error e (driver bt ...: 1 Time(s) > st 0:0:6:0: [st0] Error 1 (driver bt ...: 1 Time(s) > st 0:0:6:0: [st0] Error e (driver bt ...: 1 Time(s) > st 0:0:7:0: [st0] Error 1 (driver bt ...: 1 Time(s) > st 0:0:7:0: [st0] Error 7 (driver bt ...: 1 Time(s) > st 0:0:7:0: [st0] Error e (driver bt ...: 1 Time(s) > > st0 is the non-rewinding tape drive. As best I know, nobody does anything > with st0 -- nst0 is what I, and amanda, use. > > This looks like there might be a bent driver. > > > Amcheck: > > slot 1: volume 'sls-9' > Will write to volume 'sls-9' in slot 1. > Writing label 'sls-9' to check writability > Volume 'sls-9' is writeable. > Server check took 25.073 seconds > (brought to you by Amanda 3.5.1) > > > I'm at a loss. Amanda worked for a decade or so with the old DLT drive, then > for 3 or 4 years with the LTO -- then started breaking in every backup. I > trusted Quantum and LSI -- I ran tar to copy my entire / directory to the > Quantum tape that came with the drive, and it worked -- that ruled out the > card and the drive, so I bought a new collection of Quantum tapes to replace > the HPs I had been using. No joy. > > Tar used a block size of 512, and amanda uses 32768. I thought the block > size might be the problem, so I looked on the 'Net and found a little bit > about changing the size, but I already knew how to do that in mt. There was > nothing about changing it in amanda. Besides, it'd been working for years at >
Re: tape problem
On Fri, 03 Dec 2021 23:18:08 + ghe2001 wrote: > Any ideas, explanations, fixes? Maybe it's time to move to Bullseye? Short of that, try a kernel from backports. -- Does anybody read signatures any more? https://charlescurley.com https://charlescurley.com/blog/
Re: tape problem
On Wednesday 29 October 2003 01:26, Yogish wrote: Hi Here are the results for the commands mt -f /dev/nst0 rewind No error, prompt was returned after the command expected dd if=/dev/st0 count=1 No output on the screen? Odd. Amanda Tape Server Host Check - ERROR: running as user root instead of amanda 1+0 records in 1+0 records out su amanda -c amcheck /yourconfigname/ Amanda Tape Server Host Check - Holding disk /home/amanda: 20204072 KB disk space available, that's plenty ERROR: cannot overwrite active tape normal02 (expecting a new tape) NOTE: skipping tape-writable test Server check took 14.131 seconds Amanda Backup Client Hosts Check Client check: 1 host checked in 0.010 seconds, 0 problems found (brought to you by Amanda 2.4.2p2) So the above named tape 1. contains a label (why didn't the dd line above show it?), and 2. its been written to according to amanda. It cannot then be reused until the tapelist has been traversed and it comes around again. I have also attached amanda.conf disklist for the configuration I don't see anything wrong with the disklist unless it might be a missing spindle number, which would make amanda think they are two different drives rather than what appears to be 2 partitions on the same drive. If the latter case, that could lead to some thrashing of the drive as it seeks back and forth between the two partition. Using identical, positive spindle numbers as the last item on each line will cause the accesses to be serialized, with the second one waiting till the first is done. Actually its marginally faster, and easier on the drive. Your amanda.conf it appears has been fixed up a little, and now contains quasi-usable numbers. I'd question the default etimeout, thats a bit short on time to do a full estimate on todays multigigabyte partitions. Off the top of my head, I think I've got that set for 1200 seconds, or 20 minutes as I have a couple of 40+ gig partitions here. Your second partition could be timing out because of this. I note you do not have a dtimeout specified, the default IIRC is 600 seconds, and ISTR I've got that setup for 1800 or 30 minutes, one of my client DLE's is over 7Gb, set for compress-client-best to make it fit, a killer cpu hog on a 500mhz machine. Holding disk. Can of worms here, you have an 80Gb tape/drive, but only 10Gb usable. This probably needs to be jacked up to at least 2x the largest DLE else the holding disk will be skipped for larger DLE's, leading (possibly, certainly if any SW compression is being used) to the shoe-shining of the drive as it underflows, runs out of data, stops and backs up to restart when the buffer fills up again. Hard on the drive, tape, and a bit of a time waster. They are much happier when they can start up and run non-stop till the tape has been written. And last, a dumpcycle of 7, runspercycle of 5, really needs at least 10 tapes in the tapecycle to prevent over-writing your last good full with a potentially bad new one *IF* you weren't specifying always-full in the disklist, but were making use of amandas ability to spread out the fulls mixed with incrementals by auto-adjusting the schedule on a day by day basis. However thats not going to be really usable until you get a lot more DLE's for amanda to juggle, 2 doesn't give her much dancing room. Now, if TPTB are married to the always-full idea, then I'd reset things in your amanda.conf to: dumpcycle 1 day runspercycle 1 runtapes 1 That way, your 6 tapes won't seem like such a limitation as each one will in fact be a full backup and amanda will know thats the case from those settings. If you have the tape capacity, thats nice, but IMO wastes a lot of amanda's intelligent talent. I've overdone it a bit here, forced by having such a small drive, as dumpcycle and runspercycle are both set to 8, and I have 28 tapes in the tapecycle. So my safety ratio is 3.5 sets of fulls on hand at any one time. But then my tapes are both small (4Gb uncompressed DDS2's), not as dependable, and dirt cheap ($30 or less for a 10 pack) on ebay. Now are you beginning to understand amanda? [snip old msg] -- Cheers, Gene AMD [EMAIL PROTECTED] 320M [EMAIL PROTECTED] 512M 99.27% setiathome rank, not too shabby for a WV hillbilly Yahoo.com attornies please note, additions to this message by Gene Heskett are: Copyright 2003 by Maurice Eugene Heskett, all rights reserved.
Re: tape problem
Hi Gene Thanks for all the help. Actually it did give some error for the second command which I had pasted. dd if=/dev/st0 count=1 No output on the screen? Odd. Amanda Tape Server Host Check - ERROR: running as user root instead of amanda 1+0 records in 1+0 records out This was the error. So how do I configure amanda.conf and the disklist such that it does incremental backup during monday through thursday and do full backup on friday. Also we have 6 tapes, out which tape 5 and 6 are rotated on alternate weeks. So How do I configure this. I apprecaite your help greatly. Regards Yogish - Original Message - From: Gene Heskett [EMAIL PROTECTED] To: Yogish [EMAIL PROTECTED] Cc: [EMAIL PROTECTED] Sent: Wednesday, October 29, 2003 1:56 AM Subject: Re: tape problem On Wednesday 29 October 2003 01:26, Yogish wrote: Hi Here are the results for the commands mt -f /dev/nst0 rewind No error, prompt was returned after the command expected dd if=/dev/st0 count=1 No output on the screen? Odd. Amanda Tape Server Host Check - ERROR: running as user root instead of amanda 1+0 records in 1+0 records out su amanda -c amcheck /yourconfigname/ Amanda Tape Server Host Check - Holding disk /home/amanda: 20204072 KB disk space available, that's plenty ERROR: cannot overwrite active tape normal02 (expecting a new tape) NOTE: skipping tape-writable test Server check took 14.131 seconds Amanda Backup Client Hosts Check Client check: 1 host checked in 0.010 seconds, 0 problems found (brought to you by Amanda 2.4.2p2) So the above named tape 1. contains a label (why didn't the dd line above show it?), and 2. its been written to according to amanda. It cannot then be reused until the tapelist has been traversed and it comes around again. I have also attached amanda.conf disklist for the configuration I don't see anything wrong with the disklist unless it might be a missing spindle number, which would make amanda think they are two different drives rather than what appears to be 2 partitions on the same drive. If the latter case, that could lead to some thrashing of the drive as it seeks back and forth between the two partition. Using identical, positive spindle numbers as the last item on each line will cause the accesses to be serialized, with the second one waiting till the first is done. Actually its marginally faster, and easier on the drive. Your amanda.conf it appears has been fixed up a little, and now contains quasi-usable numbers. I'd question the default etimeout, thats a bit short on time to do a full estimate on todays multigigabyte partitions. Off the top of my head, I think I've got that set for 1200 seconds, or 20 minutes as I have a couple of 40+ gig partitions here. Your second partition could be timing out because of this. I note you do not have a dtimeout specified, the default IIRC is 600 seconds, and ISTR I've got that setup for 1800 or 30 minutes, one of my client DLE's is over 7Gb, set for compress-client-best to make it fit, a killer cpu hog on a 500mhz machine. Holding disk. Can of worms here, you have an 80Gb tape/drive, but only 10Gb usable. This probably needs to be jacked up to at least 2x the largest DLE else the holding disk will be skipped for larger DLE's, leading (possibly, certainly if any SW compression is being used) to the shoe-shining of the drive as it underflows, runs out of data, stops and backs up to restart when the buffer fills up again. Hard on the drive, tape, and a bit of a time waster. They are much happier when they can start up and run non-stop till the tape has been written. And last, a dumpcycle of 7, runspercycle of 5, really needs at least 10 tapes in the tapecycle to prevent over-writing your last good full with a potentially bad new one *IF* you weren't specifying always-full in the disklist, but were making use of amandas ability to spread out the fulls mixed with incrementals by auto-adjusting the schedule on a day by day basis. However thats not going to be really usable until you get a lot more DLE's for amanda to juggle, 2 doesn't give her much dancing room. Now, if TPTB are married to the always-full idea, then I'd reset things in your amanda.conf to: dumpcycle 1 day runspercycle 1 runtapes 1 That way, your 6 tapes won't seem like such a limitation as each one will in fact be a full backup and amanda will know thats the case from those settings. If you have the tape capacity, thats nice, but IMO wastes a lot of amanda's intelligent talent. I've overdone it a bit here, forced by having such a small drive, as dumpcycle and runspercycle are both set to 8, and I have 28 tapes in the tapecycle. So my safety ratio is 3.5 sets of fulls on hand at any one time. But then my tapes are both small (4Gb
Re: tape problem
On Wed, Oct 29, 2003 at 06:56:59PM -0700, Yogish wrote: Hi Gene Thanks for all the help. Actually it did give some error for the second command which I had pasted. dd if=/dev/st0 count=1 No output on the screen? Odd. Amanda Tape Server Host Check - ERROR: running as user root instead of amanda 1+0 records in 1+0 records out This was the error. That is not an error. That is the normal report that dd gives at the end of its run. It read one 512 byte record from the input file (/dev/st0) and wrote one 512 byte record to the output file (stdout). So how do I configure amanda.conf and the disklist such that it does incremental backup during monday through thursday and do full backup on friday. Also we have 6 tapes, out which tape 5 and 6 are rotated on alternate weeks. So How do I configure this. You are attempting to use a screwdriver to attach a nail. I.e. wrong tool for the job. You want to bend amanda away from how it wants to do the job it was designed to do so that it fits your preconceived idea of how to do backups. Not that it can't be used your way. But you will have to jump through hoops to do it. A scan of the mailing list archives for the other 20+ times it has come up this year might be in order. jon -- Jon H. LaBadie [EMAIL PROTECTED] JG Computing 4455 Province Line Road(609) 252-0159 Princeton, NJ 08540-4322 (609) 683-7220 (fax)
Re: tape problem
On Wednesday 29 October 2003 20:56, Yogish wrote: Hi Gene Thanks for all the help. Actually it did give some error for the second command which I had pasted. dd if=/dev/st0 count=1 I was refering to this^^ command when I said Odd No output on the screen? Odd. Amanda Tape Server Host Check - ERROR: running as user root instead of amanda 1+0 records in 1+0 records out This was the error. So how do I configure amanda.conf and the disklist such that it does incremental backup during monday through thursday and do full backup on friday. Also we have 6 tapes, out which tape 5 and 6 are rotated on alternate weeks. So How do I configure this. I apprecaite your help greatly. Regards Yogish The actual scheduleing of the backups is done by the user amandas crontab configuration. I forget which of the variables is the day of the week, but you replace the * in that position with a 1,2,3,4,5 and it will be run 5 times a week, sunday being 0 IIRC. See man crontab for those sparse details. amanda.conf then should define dumpcycle as 1, runspercycle as 1, and tapecycle as 6. I sent another message describing that I think. This should do fulls every nite for 5 nights a week. Amanda doesn't really have the facilities to do a backup schedule that fits a humans idea of a regular schedule, at least not without a lot of arm twisting. However, dumpcycle is amandas means of guaranteeing you that a full backup of everything WILL be done within that 7 day time period unless you have small tapes and it just won't all fit. To do that, go ahead and set dumpcycle to 1 week or 7 days, and runspercycle to 5, so amanda knows she has only 5 actual runs to get this full backup of everything done in. Amanda will then do an initial full on anything not already covered, and incrementals on the other days. But she will often do a shorter cycle until she has arrived at a working schedule. That may take several tapecycles before a 'balanced' schedule has been acheived. - Original Message - From: Gene Heskett [EMAIL PROTECTED] To: Yogish [EMAIL PROTECTED] Cc: [EMAIL PROTECTED] Sent: Wednesday, October 29, 2003 1:56 AM Subject: Re: tape problem On Wednesday 29 October 2003 01:26, Yogish wrote: Hi Here are the results for the commands mt -f /dev/nst0 rewind No error, prompt was returned after the command expected dd if=/dev/st0 count=1 No output on the screen? Odd. Amanda Tape Server Host Check - ERROR: running as user root instead of amanda 1+0 records in 1+0 records out su amanda -c amcheck /yourconfigname/ Amanda Tape Server Host Check - Holding disk /home/amanda: 20204072 KB disk space available, that's plenty ERROR: cannot overwrite active tape normal02 (expecting a new tape) NOTE: skipping tape-writable test Server check took 14.131 seconds Amanda Backup Client Hosts Check Client check: 1 host checked in 0.010 seconds, 0 problems found (brought to you by Amanda 2.4.2p2) So the above named tape 1. contains a label (why didn't the dd line above show it?), and 2. its been written to according to amanda. It cannot then be reused until the tapelist has been traversed and it comes around again. I have also attached amanda.conf disklist for the configuration I don't see anything wrong with the disklist unless it might be a missing spindle number, which would make amanda think they are two different drives rather than what appears to be 2 partitions on the same drive. If the latter case, that could lead to some thrashing of the drive as it seeks back and forth between the two partition. Using identical, positive spindle numbers as the last item on each line will cause the accesses to be serialized, with the second one waiting till the first is done. Actually its marginally faster, and easier on the drive. Your amanda.conf it appears has been fixed up a little, and now contains quasi-usable numbers. I'd question the default etimeout, thats a bit short on time to do a full estimate on todays multigigabyte partitions. Off the top of my head, I think I've got that set for 1200 seconds, or 20 minutes as I have a couple of 40+ gig partitions here. Your second partition could be timing out because of this. I note you do not have a dtimeout specified, the default IIRC is 600 seconds, and ISTR I've got that setup for 1800 or 30 minutes, one of my client DLE's is over 7Gb, set for compress-client-best to make it fit, a killer cpu hog on a 500mhz machine. Holding disk. Can of worms here, you have an 80Gb tape/drive, but only 10Gb usable. This probably needs to be jacked up to at least 2x the largest DLE else the holding disk will be skipped for larger DLE's, leading (possibly, certainly if any SW compression is being used) to the shoe-shining of the drive as it underflows, runs out of data
Re: tape problem
Yogish wrote: Hi Paul and Scheeder Here is the error it is giving me when I try to run amdump. It is not specific to friday but it happens everyday, So I am not sure what to do. I have attached the lastest amanda.conf and disklist files. In your current amanda.conf is still an error (runspercycle 5 weeks is interpreted as 35), that I pointed out previously; I didn't read any further to see other errors anymore. Why don't you just do what experienced users on this list propose? *** A TAPE ERROR OCCURRED: [reading label: Input/output error]. Some dumps may have been left in the holding disk. Run amflush to flush them to tape. The next tape Amanda expects to use is: a new tape. FAILURE AND STRANGE DUMP SUMMARY: brokerserv sda6 lev 0 FAILED [can't switch to incremental dump] brokerserv sda3 lev 0 FAILED [can't switch to incremental dump] Please tell how I can force amanda to do a dump on the existing tape irrespective of the tape number or label. Amanda NEEDS a label to verify. Just label the tape, *please*! e.g. why don't you verify if there is a label on the tape with the command Christophe Scheeder explained? And if necessary just relabel the tape with amlabel -f ... just as explained in the same mail. Repeating the same thing just makes me tired, or ignore further requests for help... -- Paul @ Home
Re: tape problem
On Monday 27 October 2003 02:02, Yogish wrote: Hi Paul and Scheeder Here is the error it is giving me when I try to run amdump. It is not specific to friday but it happens everyday, So I am not sure what to do. I have attached the lastest amanda.conf and disklist files. *** A TAPE ERROR OCCURRED: [reading label: Input/output error]. Some dumps may have been left in the holding disk. Run amflush to flush them to tape. The next tape Amanda expects to use is: a new tape. FAILURE AND STRANGE DUMP SUMMARY: brokerserv sda6 lev 0 FAILED [can't switch to incremental dump] brokerserv sda3 lev 0 FAILED [can't switch to incremental dump] Please tell how I can force amanda to do a dump on the existing tape irrespective of the tape number or label. Regards Yogish You first have to get rid of the i/o error above, Yogish. This is the reason it cannot access the tape, and must be fixed before anything else is going to work. To give an example, I had an amanda error last night, an i/o error, so it left the backup in the holding disk. I have my tapes all setup for a blocksize of 32k, but apparently that didn't take when I did a 5 minute boot to 2.6.0-test9 to see what may, or may not be working. Amcheck couldn't access the drive at all, oss-install wouldn't, and nvidia's drivers wouldn't, so I had no X and rebooted back to 2.4.23-pre8. So to restore tape access, I had to redo just now, by hand, the pair of mt commands that set the working blocksize, and amflush is running ok as I type this. I have those commands in /etc/rc.d/rc.local also, so I'm puzzled as to why that wasn't self-repairing on the reboot. So what blocksize are you using? The default for most drives is 512 bytes, which is also the default size for the mt driver. You will have an i/o error if they do not match. That of course is just one possible scenario that can cause an i/o error. [snip, with comment] Let me reinforce the requirement that all operations on the drive performed by amanda must be done using the non-rewinding device, I believe /dev/nst0 in your case. Amanda will rewind the tape when amanda wants it rewound, and expects that. To have it done automaticly by closing the device, as when using /dev/st0, will thoroughly fubar amandas carefull tape handling. -- Cheers, Gene AMD [EMAIL PROTECTED] 320M [EMAIL PROTECTED] 512M 99.27% setiathome rank, not too shabby for a WV hillbilly Yahoo.com attornies please note, additions to this message by Gene Heskett are: Copyright 2003 by Maurice Eugene Heskett, all rights reserved.
Re: tape problem
Hi I am facing two problems. first of all I need to force the label for the inserted tape and only then dump works. However There is seems to be a problem with the amdump. It cannot dump sda3 but can dump sda6 successfully.Here is the error on the log file for amdump. I have deleted the weeks from runspercycle as told. I am not able to figure out the problem. I would appreciate if you would help me with this. START planner date 20031027 INFO planner Forcing full dump of brokerserver:sda3 as directed. ERROR taper no-tape [reading label: Input/output error] FINISH planner date 20031027 STATS driver startup time 106.953 FAIL driver brokerserver sda3 0 [can't switch to incremental dump] SUCCESS dumper brokerserver sda6 20031027 1 [sec 684.154 kb 4602880 kps 6727.8 orig-kb 6113730] FINISH driver date 20031027 time 791.162 Regards Yogish - Original Message - From: Paul Bijnens [EMAIL PROTECTED] To: Yogish [EMAIL PROTECTED] Cc: [EMAIL PROTECTED] Sent: Monday, October 27, 2003 1:08 AM Subject: Re: tape problem Yogish wrote: Hi Paul and Scheeder Here is the error it is giving me when I try to run amdump. It is not specific to friday but it happens everyday, So I am not sure what to do. I have attached the lastest amanda.conf and disklist files. In your current amanda.conf is still an error (runspercycle 5 weeks is interpreted as 35), that I pointed out previously; I didn't read any further to see other errors anymore. Why don't you just do what experienced users on this list propose? *** A TAPE ERROR OCCURRED: [reading label: Input/output error]. Some dumps may have been left in the holding disk. Run amflush to flush them to tape. The next tape Amanda expects to use is: a new tape. FAILURE AND STRANGE DUMP SUMMARY: brokerserv sda6 lev 0 FAILED [can't switch to incremental dump] brokerserv sda3 lev 0 FAILED [can't switch to incremental dump] Please tell how I can force amanda to do a dump on the existing tape irrespective of the tape number or label. Amanda NEEDS a label to verify. Just label the tape, *please*! e.g. why don't you verify if there is a label on the tape with the command Christophe Scheeder explained? And if necessary just relabel the tape with amlabel -f ... just as explained in the same mail. Repeating the same thing just makes me tired, or ignore further requests for help... -- Paul @ Home --- Outgoing mail is certified Virus Free. Checked by AVG anti-virus system (http://www.grisoft.com). Version: 6.0.529 / Virus Database: 324 - Release Date: 10/16/2003
Re: tape problem
On Monday 27 October 2003 22:05, Yogish wrote: Hi I am facing two problems. first of all I need to force the label for the inserted tape and only then dump works. However There is seems to be a problem with the amdump. It cannot dump sda3 but can dump sda6 successfully.Here is the error on the log file for amdump. I have deleted the weeks from runspercycle as told. I am not able to figure out the problem. I would appreciate if you would help me with this. START planner date 20031027 INFO planner Forcing full dump of brokerserver:sda3 as directed. ERROR taper no-tape [reading label: Input/output error] And you still have this error.^^ As I said before, nothing else is going to work UNTIL this is fixed. FINISH planner date 20031027 STATS driver startup time 106.953 FAIL driver brokerserver sda3 0 [can't switch to incremental dump] SUCCESS dumper brokerserver sda6 20031027 1 [sec 684.154 kb 4602880 kps 6727.8 orig-kb 6113730] FINISH driver date 20031027 time 791.162 Please clip and post the results you see on screen of these three commands, executed against the tape in the drive when amdump was last run. As root: mt -f /dev/nst0 rewind dd if=/dev/st0 count=1 su amanda -c amcheck /yourconfigname/ where /yourconfigname/ is whatever you have named it. It is the directory name where you will find your amanda.conf. Also, Please post your disklist entry for sda3, and your amanda.conf Regards Yogish [snip old pleadings to please do as we say] -- Cheers, Gene AMD [EMAIL PROTECTED] 320M [EMAIL PROTECTED] 512M 99.27% setiathome rank, not too shabby for a WV hillbilly Yahoo.com attornies please note, additions to this message by Gene Heskett are: Copyright 2003 by Maurice Eugene Heskett, all rights reserved.
Re: tape problem
Yogish wrote: Hi all I am facing problems with the tapes. Everytime I try to run amcheck,amverify or amdump it says its expecting a new tape. Here are the results it puts out when I run amcheck Amanda Tape Server Host Check - Holding disk /home/amanda: 29635568 KB disk space available, that's plenty ERROR: /dev/nst0: reading label: Input/output error (expecting a new tape) NOTE: skipping tape-writable test Server check took 0.016 seconds I strongly believe that the amanda.conf you send is not the one that is used! Or you changed it very recently. From your amanda.conf: tapedev /dev/st0 This is wrong: it is the rewinding device, and if I remember correctly I pointed out this and other errors in your amanda.conf a few weeks ago: http://groups.yahoo.com/group/amanda-users/message/46994 But the amcheck above says it uses /dev/nst0, so assume that the current amanda.conf is a different one... It could be that by using the rewinding device, you accidently overwrote the label. Was this what happened? -- Paul @ Home
Re: tape problem
On Sun, Oct 26, 2003 at 10:28:57AM -0700, Yogish wrote: Hi all I am facing problems with the tapes. Everytime I try to run amcheck,amverify or amdump it says its expecting a new tape. Here are the results it puts out when I run amcheck Amanda Tape Server Host Check - Holding disk /home/amanda: 29635568 KB disk space available, that's plenty ERROR: /dev/nst0: reading label: Input/output error (expecting a new tape) NOTE: skipping tape-writable test Server check took 0.016 seconds Have you amlabeled the tapes? The amcheck output makes it appear 'no'. How many tapes does your tapecycle say and how many have been used so far. Until you have used that many tapes it always wants a new tape. -- Jon H. LaBadie [EMAIL PROTECTED] JG Computing 4455 Province Line Road(609) 252-0159 Princeton, NJ 08540-4322 (609) 683-7220 (fax)
Re: tape problem
Hi Paul You were right in pointing out that I have used a different amanda.conf file. However the changes in this file and that file are minimal (like nst0 instead of st0). You might be right when you say by using a rewinding device, I might have overwirtten the name , I am not sure how to answer this. I tried to rename the tape by amlabel normal normal05 which should be the tape it should be expecting. When I do this it says the label is already present in the drive. How can do a manual dump for time being immediately?. IS there any way I can force amanda to do backup on the tape irrespective of its label?. I appreciate your help. Regards Yogish Yogish wrote: Hi all I am facing problems with the tapes. Everytime I try to run amcheck,amverify or amdump it says its expecting a new tape. Here are the results it puts out when I run amcheck Amanda Tape Server Host Check - Holding disk /home/amanda: 29635568 KB disk space available, that's plenty ERROR: /dev/nst0: reading label: Input/output error (expecting a new tape) NOTE: skipping tape-writable test Server check took 0.016 seconds I strongly believe that the amanda.conf you send is not the one that is used! Or you changed it very recently. From your amanda.conf: tapedev /dev/st0 This is wrong: it is the rewinding device, and if I remember correctly I pointed out this and other errors in your amanda.conf a few weeks ago: http://groups.yahoo.com/group/amanda-users/message/46994 But the amcheck above says it uses /dev/nst0, so assume that the current amanda.conf is a different one... It could be that by using the rewinding device, you accidently overwrote the label. Was this what happened? -- Paul @ Home --
Re: tape problem
Hi, it is easy for you to verify if your tapelabels are valid, do the following commands: mt rewind dd if=/dev/nst0 bs=32k count=1 this will give you the first block of the tape. if the first characters there in are something like AMANDA: TAPESTART DATE then your labels are ok. if not you'll have to relabel them. How? very easy: amlabel -f normal normal05 the -f option forces amanda to ignore its tape-database and the contents on tape an write the label anyway. Christoph Your Name schrieb: Hi Paul You were right in pointing out that I have used a different amanda.conf file. However the changes in this file and that file are minimal (like nst0 instead of st0). You might be right when you say by using a rewinding device, I might have overwirtten the name , I am not sure how to answer this. I tried to rename the tape by amlabel normal normal05 which should be the tape it should be expecting. When I do this it says the label is already present in the drive. How can do a manual dump for time being immediately?. IS there any way I can force amanda to do backup on the tape irrespective of its label?. I appreciate your help. Regards Yogish Yogish wrote: Hi all I am facing problems with the tapes. Everytime I try to run amcheck,amverify or amdump it says its expecting a new tape. Here are the results it puts out when I run amcheck Amanda Tape Server Host Check - Holding disk /home/amanda: 29635568 KB disk space available, that's plenty ERROR: /dev/nst0: reading label: Input/output error (expecting a new tape) NOTE: skipping tape-writable test Server check took 0.016 seconds I strongly believe that the amanda.conf you send is not the one that is used! Or you changed it very recently. From your amanda.conf: tapedev /dev/st0 This is wrong: it is the rewinding device, and if I remember correctly I pointed out this and other errors in your amanda.conf a few weeks ago: http://groups.yahoo.com/group/amanda-users/message/46994 But the amcheck above says it uses /dev/nst0, so assume that the current amanda.conf is a different one... It could be that by using the rewinding device, you accidently overwrote the label. Was this what happened? -- Paul @ Home
Re: tape problem
Sergio Pereira wrote: hi folks, I'm using amanda-2.4.2p2 with redhat 7.2 and I'm geting problem to setup some tapes to use. When I try to use amlabel I get a error message: amlabel: no tape online but the tape is there. any thoughts about it ? thanks in advance Sergio what is your 'tapedev' in your amanda.conf? what does 'mt -f 'entry from amanda.conf' tell' say? -- Christopher McCrory The guy that keeps the servers running [EMAIL PROTECTED] http://www.pricegrabber.com True, it returns for false, but is an even more interesting number than 0. -- Larry Wall in [EMAIL PROTECTED]
