[Bacula-users] Backing up OS X and Windows clients and keeping them awake
For those that are backing up OS X clients, you may have noticed that 10.8 Mountain Lion and later is much more militant about having machines sleep. Even if you issue a WOL packet to start a backup, they still fall back asleep pretty quickly unless there’s someone actually using the mouse and keyboard. It doesn’t really matter what you have the sleep timer set to. Apple’s strategy seems to be that programs that need to run must issue a “power assertion” to the system, indicating that the system must stay awake for them complete their task. Fortunately, it’s easy enough to issue one of these manually using the command line. In my job definition (in the Bacula director configuration) I’ve added this line to my Mac clients: Client Run Before Job = "/usr/bin/screen -d -m /usr/bin/caffeinate -i -t 3600; sleep 0;” The first part, using the “screen” command, puts the command in the background. The “caffeinate” command is new for 10.8 and later, and will keep the system awake for 3600 seconds. (Edit to suit your needs.) If you’re backing up 10.7 or earlier systems, they won’t have the caffeinate command, so trying to issue it will fail and thus your backup would fail. That’s why there’s the “sleep 0” part at the end. It ensures you’ll always exit this command with a success (0). Hopefully this will help someone else that gets stuck when a Mac backup starts but then fails, usually taking about 2 hours before the director will give up and move on. Windows 7 seems to have a similar issue. I was able to solve that by editing a registry value as below: [HKEY_LOCAL_MACHINE\System\CurrentControlSet\Control\Power\PowerSettings\238C9FA8-0AAD-41ED-83F4-97BE242C8F20\7bc4a2f9-d8fc-4469-b07b-33eb785aaca0\DefaultPowerSchemeValues\381b4222-f694-41f0-9685-ff5bb260df2e] "AcSettingIndex"=dword:0e10 [HKEY_LOCAL_MACHINE\System\CurrentControlSet\Control\Power\PowerSettings\238C9FA8-0AAD-41ED-83F4-97BE242C8F20\7bc4a2f9-d8fc-4469-b07b-33eb785aaca0\DefaultPowerSchemeValues\8c5e7fda-e8bf-4a96-9a85-a6e23a8c635c] "AcSettingIndex"=dword:0e10 [HKEY_LOCAL_MACHINE\System\CurrentControlSet\Control\Power\PowerSettings\238C9FA8-0AAD-41ED-83F4-97BE242C8F20\7bc4a2f9-d8fc-4469-b07b-33eb785aaca0\DefaultPowerSchemeValues\a1841308-3541-4fab-bc81-f71556f20b4a] "AcSettingIndex"=dword:0e10 Cheers! -Joe Rhodes smime.p7s Description: S/MIME cryptographic signature -- CenturyLink Cloud: The Leader in Enterprise Cloud Services. Learn Why More Businesses Are Choosing CenturyLink Cloud For Critical Workloads, Development Environments & Everything In Between. Get a Quote or Start a Free Trial Today. http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] Backing up OS X and Windows clients and keeping them awake
Sorry, turns out there was an error in my previous suggestion and “caffeinate” was not actually started on the client. Here’s an updated (and tested) line that does keep a 10.8 or later mac awake for 3600 seconds: Client Run Before Job = "bash -c \"/usr/bin/caffeinate -i -t 3600 &> /dev/null < /dev/null &\"" Turns out, getting a command into the background using the “Client Run Before Job” was more tricky than I realized. This command does the trick. You’ll have to adjust the time (3600) to suit your needs. (I’ve got one client that takes 4 hours to do a full backup.) It also does not fail the job if the “caffeinate” command cannot be found (10.7 or earlier clients). So it should be safe to use that command for any OS X client. Josh Fisher asked: > I take there simply is no way to prevent sleep on 10.7 and earlier? With 10.7 and earlier, things were actually simpler. You just set the sleep timer in System Preferences (or using pmset from the command line) and the Mac would stay awake for whatever you had it set for, even from a WOL packet. I’d typically set it for 2 hours on most machines. I only started to get lots of failed jobs with 10.8 and later. Thats’ where Apple got a lot more aggressive about power management. > You wouldn't happen to know how to enable WoWLAN on a Mac would you? Again, easily set in System Preferences or using pmset from the command line. If you’ve got a lot of Macs, Apple Remote Desktop is your friend here. (“Send Unix Command” to the whole fleet) Cheers! -Joe Rhodes > For those that are backing up OS X clients, you may have noticed that 10.8 > Mountain Lion and later is much more militant about having machines sleep. > Even if you issue a WOL packet to start a backup, they still fall back asleep > pretty quickly unless there?s someone actually using the mouse and keyboard. > It doesn?t really matter what you have the sleep timer set to. > > Apple?s strategy seems to be that programs that need to run must issue a > ?power assertion? to the system, indicating that the system must stay awake > for them complete their task. Fortunately, it?s easy enough to issue one of > these manually using the command line. > > In my job definition (in the Bacula director configuration) I?ve added this > line to my Mac clients: > > Client Run Before Job = "/usr/bin/screen -d -m /usr/bin/caffeinate -i -t > 3600; sleep 0;? > > The first part, using the ?screen? command, puts the command in the > background. The ?caffeinate? command is new for 10.8 and later, and will > keep the system awake for 3600 seconds. (Edit to suit your needs.) > > If you?re backing up 10.7 or earlier systems, they won?t have the caffeinate > command, so trying to issue it will fail and thus your backup would fail. > That?s why there?s the ?sleep 0? part at the end. It ensures you?ll always > exit this command with a success (0). > > Hopefully this will help someone else that gets stuck when a Mac backup > starts but then fails, usually taking about 2 hours before the director will > give up and move on. > > Windows 7 seems to have a similar issue. I was able to solve that by editing > a registry value as below: > > > [HKEY_LOCAL_MACHINE\System\CurrentControlSet\Control\Power\PowerSettings\238C9FA8-0AAD-41ED-83F4-97BE242C8F20\7bc4a2f9-d8fc-4469-b07b-33eb785aaca0\DefaultPowerSchemeValues\381b4222-f694-41f0-9685-ff5bb260df2e] > "AcSettingIndex"=dword:0e10 > > [HKEY_LOCAL_MACHINE\System\CurrentControlSet\Control\Power\PowerSettings\238C9FA8-0AAD-41ED-83F4-97BE242C8F20\7bc4a2f9-d8fc-4469-b07b-33eb785aaca0\DefaultPowerSchemeValues\8c5e7fda-e8bf-4a96-9a85-a6e23a8c635c] > "AcSettingIndex"=dword:0e10 > > [HKEY_LOCAL_MACHINE\System\CurrentControlSet\Control\Power\PowerSettings\238C9FA8-0AAD-41ED-83F4-97BE242C8F20\7bc4a2f9-d8fc-4469-b07b-33eb785aaca0\DefaultPowerSchemeValues\a1841308-3541-4fab-bc81-f71556f20b4a] > "AcSettingIndex"=dword:0e10 > > Cheers! > -Joe Rhodes -- CenturyLink Cloud: The Leader in Enterprise Cloud Services. Learn Why More Businesses Are Choosing CenturyLink Cloud For Critical Workloads, Development Environments & Everything In Between. Get a Quote or Start a Free Trial Today. http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Backing up OS X and Windows clients and
Well, as it turns out, there’s an slightly better way to do this, at least for now. >> > > I see. With 10.7 and earlier there is no way to "temporarily" adjust the > sleep timer. If a permanent long sleep timer is not desired, then it > would require the RunBefore script to set the sleep timer with pmset and > a RunAfter job to set it back to normal following the backup. My Mac > clients are in and out of the office seemingly at random. It is not > uncommon for the users to leave in the middle of a backup. Up grading to > 10.8 and using caffeinate seems to be my best option, else they will be > left with a long sleep timer while traveling should they leave in the > middle of a backup. It is a real pain trying to backup these "road > warrior" clients. > There’s an option to pmset called “noidle” that will keep the mac awake as long as the command is running. It runs in the foreground. So here’s what I do: Client Run Before Job = "bash -c \"/usr/bin/pmset noidle &> /dev/null < /dev/null &\"" Client Run After Job = "/usr/bin/killall pmset" The first command will run pmset with the “noidle” option in the background. This will keep the Mac awake. The second one goes through and kills the pmset command after the backup is done, thus allowing it to fall asleep per it’s schedule. That should take care of it for any recent OS X client, including 10.6 and up. Not sure about 10.5 and earlier. It should also handle backup jobs of arbitrary time. Also, if your Macs leave in the middle of a backup operation, they’ll have that running. A restart will correct that. So perhaps the “caffeinate” command is better for those folks. At least that will expire on it’s own. Note that the “noidle” option is deprecated as of 10.8 or 10.9 and Apple officially recommends using caffeinate instead. So we might have to rethink this sometime with the next release of OS X. Sorry for all the back and forth on this. It’s something I’ve been wrestling with for a while. I think this will be my final answer. :) Cheers! -Joe Rhodes >> >>> You wouldn't happen to know how to enable WoWLAN on a Mac would you? >> >> Again, easily set in System Preferences or using pmset from the command >> line. If you?ve got a lot of Macs, Apple Remote Desktop is your friend >> here. (?Send Unix Command? to the whole fleet) > > Thanks, Joe. > >> >> Cheers! >> -Joe Rhodes >> >> >>> For those that are backing up OS X clients, you may have noticed that 10.8 >>> Mountain Lion and later is much more militant about having machines sleep. >>> Even if you issue a WOL packet to start a backup, they still fall back >>> asleep pretty quickly unless there?s someone actually using the mouse and >>> keyboard. It doesn?t really matter what you have the sleep timer set to. >>> >>> Apple?s strategy seems to be that programs that need to run must issue a >>> ?power assertion? to the system, indicating that the system must stay awake >>> for them complete their task. Fortunately, it?s easy enough to issue one >>> of these manually using the command line. >>> >>> In my job definition (in the Bacula director configuration) I?ve added >>> this line to my Mac clients: >>> >>> Client Run Before Job = "/usr/bin/screen -d -m /usr/bin/caffeinate -i -t >>> 3600; sleep 0;? >>> >>> The first part, using the ?screen? command, puts the command in the >>> background. The ?caffeinate? command is new for 10.8 and later, and will >>> keep the system awake for 3600 seconds. (Edit to suit your needs.) >>> >>> If you?re backing up 10.7 or earlier systems, they won?t have the >>> caffeinate command, so trying to issue it will fail and thus your backup >>> would fail. That?s why there?s the ?sleep 0? part at the end. It ensures >>> you?ll always exit this command with a success (0). >>> >>> Hopefully this will help someone else that gets stuck when a Mac backup >>> starts but then fails, usually taking about 2 hours before the director >>> will give up and move on. >>> >>> Windows 7 seems to have a similar issue. I was able to solve that by >>> editing a registry value as below: >>> >>> >>> [HKEY_LOCAL_MACHINE\System\CurrentControlSet\Control\Power\PowerSettings\238C9FA8-0AAD-41ED-83F4-97BE242C8F20\7bc4a2f9-d8fc-4469-b07b-33eb785aaca0\DefaultPowerSchemeValues\381b4222-f694-41f0-9685-ff5bb260df2e] >>> "AcSettingIndex"=dword:0e10 >>> >>> [HKEY_LOCAL_MACHINE
Re: [Bacula-users] Backing up OS X and Windows clients and keeping them awake
Kern: I no very little about programming, but here’s the documentation for creating and releasing a power assertion on OS X. There is a call “IOPMAssertionCreate” which is available in 10.5, then deprecated in 10.6. Seems it was replaced with “IOPMAssertionCreateWithDescription” which came in 10.7. https://developer.apple.com/library/mac/documentation/IOKit/Reference/IOPMLib_header_reference/Reference/reference.html#//apple_ref/doc/uid/TP40012430 I don’t know if that will help any or not. It would be pretty awesome to see in Bacula. Love that Windows will be getting that feature built in! Cheers! -Joe > Hello, > > The next Windows version to be released by Bacula Systems roughly in > June -- it will > be version 8.0 tells the Windows OS not to suspend the SD during a job. > > For OSX, I don't know if an OS API exists to do this -- on Linux, it > doesn't seem > to have one, which means that it is not so simple. If anyone has some > simple > OS API call for Linux or Mac OSX that will do this, please let me know. > > Best regards, > Kern > > On 01/27/2014 04:50 PM, Josh Fisher wrote: >> On 1/21/2014 6:25 PM, Joe Rhodes wrote: >>> Sorry, turns out there was an error in my previous suggestion and >>> ?caffeinate? was not actually started on the client. Here?s an updated >>> (and tested) line that does keep a 10.8 or later mac awake for 3600 seconds: >>> >>> Client Run Before Job = "bash -c \"/usr/bin/caffeinate -i -t 3600 &> >>> /dev/null < /dev/null &\"" >>> >>> >>> Turns out, getting a command into the background using the ?Client Run >>> Before Job? was more tricky than I realized. This command does the trick. >>> You?ll have to adjust the time (3600) to suit your needs. (I?ve got one >>> client that takes 4 hours to do a full backup.) >>> >>> It also does not fail the job if the ?caffeinate? command cannot be found >>> (10.7 or earlier clients). So it should be safe to use that command for >>> any OS X client. >>> >>> >>> Josh Fisher asked: >>> >>>> I take there simply is no way to prevent sleep on 10.7 and earlier? >>> With 10.7 and earlier, things were actually simpler. You just set the >>> sleep timer in System Preferences (or using pmset from the command line) >>> and the Mac would stay awake for whatever you had it set for, even from a >>> WOL packet. I?d typically set it for 2 hours on most machines. I only >>> started to get lots of failed jobs with 10.8 and later. Thats? where Apple >>> got a lot more aggressive about power management. >> I see. With 10.7 and earlier there is no way to "temporarily" adjust the >> sleep timer. If a permanent long sleep timer is not desired, then it >> would require the RunBefore script to set the sleep timer with pmset and >> a RunAfter job to set it back to normal following the backup. My Mac >> clients are in and out of the office seemingly at random. It is not >> uncommon for the users to leave in the middle of a backup. Up grading to >> 10.8 and using caffeinate seems to be my best option, else they will be >> left with a long sleep timer while traveling should they leave in the >> middle of a backup. It is a real pain trying to backup these "road >> warrior" clients. >> >>>> You wouldn't happen to know how to enable WoWLAN on a Mac would you? >>> Again, easily set in System Preferences or using pmset from the command >>> line. If you?ve got a lot of Macs, Apple Remote Desktop is your friend >>> here. (?Send Unix Command? to the whole fleet) >> Thanks, Joe. >> >>> Cheers! >>> -Joe Rhodes >>> >>> >>>> For those that are backing up OS X clients, you may have noticed that 10.8 >>>> Mountain Lion and later is much more militant about having machines sleep. >>>> Even if you issue a WOL packet to start a backup, they still fall back >>>> asleep pretty quickly unless there?s someone actually using the mouse and >>>> keyboard. It doesn?t really matter what you have the sleep timer set to. >>>> >>>> Apple?s strategy seems to be that programs that need to run must issue a >>>> ?power assertion? to the system, indicating that the system must stay >>>> awake for them complete their task. Fortunately, it?s easy enough to >>>> issue one of these manually using the command line. >>>> >>>> In my job definition (in the Bacula director configura
Re: [Bacula-users] Backing up OS X and Windows clients and keeping them awake
Peter: If you run caffinate on 10.9, it runs in the foreground. Same with “pmset noidle” command. You have to take some measures to get it to run in the background on the client. This took a little fooling around on my part before I got it right. Here’s what I use in my bacular-dir.conf: Client Run Before Job = "bash -c \"/usr/bin/pmset noidle &> /dev/null < /dev/null &\”” Client Run After Job = "/usr/bin/killall pmset" That should put the pmset command into the background, then kill it after the job has been run. Watch the double quotes. They have to be escaped. It’s a little tricky to get bacula to run a “Client Run Before” command into the background. You have to redirect it’s input and output. If I’m seeing it correctly, I think you’ve got a space between the & and the > right after “noidle". They should be right on top of each other, no spaces. See section 3.6 of this how-to: http://tldp.org/HOWTO/Bash-Prog-Intro-HOWTO-3.html Cheers! -Joe On Feb 5, 2014, at 4:56 PM, Peter Wood wrote: > > On Tue, Jan 21, 2014 at 3:25 PM, Joe Rhodes wrote: > Sorry, turns out there was an error in my previous suggestion and > “caffeinate” was not actually started on the client. Here’s an updated (and > tested) line that does keep a 10.8 or later mac awake for 3600 seconds: > > Client Run Before Job = "bash -c \"/usr/bin/caffeinate -i -t 3600 &> > /dev/null < /dev/null &\"" > > > With Bacula 5.2.13 running on CentOS6.4 the command doesn't go in on OS X > 10.9.1 client. > > On the OS X system I see it running: > sdfiler:~ root# ps -ef |grep caffeinate > 0 48955 1 0 10:11AM ?? 0:00.01 /usr/bin/caffeinate -i -t > 21600 > > ... but the backup stops waiting for the command to complete before > continuing. If I kill the caffeinate process the backup continues instantly. > > If I don't kill the caffeinate process the connection stays idle for 15min. > and the OS X closes it. > > Here is the output: > 05-Feb 10:11 tzbackup-dir JobId 3965: No prior Full backup Job record found. > 05-Feb 10:11 tzbackup-dir JobId 3965: No prior or suitable Full backup found > in catalog. Doing FULL backup. > 05-Feb 10:11 tzbackup-dir JobId 3965: shell command: run BeforeJob > "/usr/local/sbin/baculaFirewall open 192.168.10.40" > 05-Feb 10:11 tzbackup-dir JobId 3965: Start Backup JobId 3965, > Job=sdfiler.2014-02-05_10.11.09_24 > 05-Feb 10:11 tzbackup-dir JobId 3965: Using Device "FileStorage" to write. > 05-Feb 10:11 sdfiler-fd JobId 3965: shell command: run ClientRunBeforeJob > "bash -c "/usr/bin/caffeinate -i -t 21600 & > /dev/null < /dev/null &"" > 05-Feb 10:26 tzbackup-dir JobId 3965: Fatal error: Socket error on > ClientRunBeforeJob command: ERR=Connection reset by peer > 05-Feb 10:26 tzbackup-dir JobId 3965: Fatal error: Client "sdfiler-fd" > RunScript failed. > 05-Feb 10:26 tzbackup-dir JobId 3965: Fatal error: Network error with FD > during Backup: ERR=Connection reset by peer > 05-Feb 10:26 tzbackup-dir JobId 3965: Fatal error: No Job status returned > from FD. > 05-Feb 10:26 tzbackup-dir JobId 3965: Error: Bacula tzbackup-dir 5.2.13 > (19Jan13): > Build OS: x86_64-unknown-linux-gnu redhat > JobId: 3965 > Job:sdfiler.2014-02-05_10.11.09_24 > Backup Level: Full (upgraded from Incremental) > Client: "sdfiler-fd" 5.2.13 (19Jan13) > i386-apple-darwin13.0.0,osx,13.0.0 > FileSet:"sdfiler-files" 2014-02-05 09:57:47 > Pool: "FullFilePool" (From Job FullPool override) > Catalog:"MyCatalog" (From Client resource) > Storage:"File" (From Job resource) > Scheduled time: 05-Feb-2014 10:10:46 > Start time: 05-Feb-2014 10:11:11 > End time: 05-Feb-2014 10:26:14 > Elapsed time: 15 mins 3 secs > Priority: 10 > FD Files Written: 0 > SD Files Written: 0 > FD Bytes Written: 0 (0 B) > SD Bytes Written: 0 (0 B) > Rate: 0.0 KB/s > Software Compression: None > VSS:no > Encryption: no > Accurate: no > Volume name(s): > Volume Session Id: 688 > Volume Session Time:1389056422 > Last Volume Bytes: 78,232,168,558 (78.23 GB) > Non-fatal FD errors:1 > SD Errors: 0 > FD termination status: Error > SD termination status: OK > Termination:*** Backup Error *** >
[Bacula-users] Bacula 7 will only see the first storage resource in bacula-dir.conf
Hey everyone!
I’m running Bacula 7.0.4 on Ubuntu 14.04. I’m backing up to two different
devices. One is a hot-swap hard drive bay which I’m treating as a changer
using the vchanger program. These drives get changed out and rotated off
site. The other is a fixed drive. The two devices will use two distinct
pools.
The issue I’m having is that no matter what I do, Bacula seems to see the FIRST
storage resource only. This ultimately causes problems because bacula will
look to the wrong device for volumes associated with the pool the job calls for.
I have a similar setup on a Bacula 5.2 setup that is working great. So I’m not
sure what I’m doing wrong. Config files look the same between the two. If
anyone has some suggestions, I’d be eternally grateful.
The stanza’s from bacula-dir.com:
#
#
# Storage
#
#
# Definition of file Virtual Autochanger device
Storage {
Name = removeable-drives
Address = 10.11.1.2 # N.B. Use a fully qualified name here
SDPort = 9103
Password = "RoundHouseBackup"
Device = ExternalDrives
Media Type = File
Maximum Concurrent Jobs = 1# run up to 10 jobs a the same time
}
Storage {
Name = LocalStorage
Address = 10.11.1.2 # N.B. Use a fully qualified name here
SDPort = 9103
Password = "RoundHouseBackup"
Device = LocalStorage
Media Type = File1
Maximum Concurrent Jobs = 1# run up to 10 jobs a the same time
}
And the stanza’s from my bacula-sd.conf:
# Define a Virtual autochanger
#
Autochanger {
Name = ExternalDrives
Device = ExternalDrives-Drive1
Changer Command = "/usr/local/bin/vchanger -u root %c %o %S %a %d"
Changer Device = "/etc/bacula/vchanger1.conf"
}
Device {
Name = ExternalDrives-Drive1
DriveIndex = 0
Autochanger = yes
DeviceType = File
Media Type = File
Archive Device = /var/lib/bacula/removeable-drives/0/drive0
#LabelMedia = yes; # lets Bacula label unlabeled media
Random Access = Yes;
AutomaticMount = yes; # when device opened, read it
RemovableMedia = no;
AlwaysOpen = no;
Maximum Concurrent Jobs = 5;
Maximum Volume Size = 476250 MB; # with 8 volumes per drive, this should
leave about
# 4,800 MB free for other things, or about
0.125%
}
Device {
Name = LocalStorage
Device Type = File
Media Type = File1
Archive Device = /mnt/localstorage/
LabelMedia = yes; # lets Bacula label unlabeled media
Random Access = Yes;
AutomaticMount = yes; # when device opened, read it
RemovableMedia = no;
AlwaysOpen = no;
Maximum Volume Size = 237500 MB
}
When I execute “status sd” in bconsole, I get this:
*status sd
Automatically selected Storage: removeable-drives
Connecting to Storage daemon removeable-drives at 10.11.1.2:9103
mail.roundhouse.local-sd Version: 7.0.4 (04 June 2014) x86_64-unknown-linux-gnu
ubuntu 14.04
Daemon started 15-Jul-14 21:55. Jobs: run=0, running=0.
Heap: heap=270,336 smbytes=198,653 max_bytes=354,151 bufs=88 max_bufs=90
Sizes: boffset_t=8 size_t=8 int32_t=4 int64_t=8 mode=0,0
Running Jobs:
No Jobs running.
…
Device status:
Autochanger "ExternalDrives" with devices:
"ExternalDrives-Drive1" (/var/lib/bacula/removeable-drives/0/drive0)
Device "ExternalDrives-Drive1" (/var/lib/bacula/removeable-drives/0/drive0) is
not open.
Drive 0 is not loaded.
==
Device "LocalStorage" (/mnt/localstorage/) is not open.
==
Used Volume status:
If I swap the order of the two “Storage” stanza’s in bacula-dir.cong, I get
this:
*status sd
Automatically selected Storage: LocalStorage
Connecting to Storage daemon LocalStorage at 10.11.1.2:9103
mail.roundhouse.local-sd Version: 7.0.4 (04 June 2014) x86_64-unknown-linux-gnu
ubuntu 14.04
Daemon started 15-Jul-14 21:55. Jobs: run=0, running=0.
Heap: heap=270,336 smbytes=199,230 max_bytes=354,151 bufs=89 max_bufs=91
Sizes: boffset_t=8 size_t=8 int32_t=4 int64_t=8 mode=0,0
Running Jobs:
No Jobs running.
...
Device status:
Autochanger "ExternalDrives" with devices:
"ExternalDrives-Drive1" (/var/lib/bacula/removeable-drives/0/drive0)
Device "ExternalDrives-Drive1" (/var/lib/bacula/removeable-drives/0/drive0) is
not open.
Drive 0 is not loaded.
==
Device "LocalStorage" (/mnt/localstorage/) is not open.
==
Used Volume status:
Notice how the “Automatically selected Storage:” line changes? It just picks
the first one and seems to ignore the other. On the 5.2 install, I get a sub
menu asking which Storage I want to get status on:
*status sd
The defined Storage resources are:
1: File
2: LocalStorage
Select Storage resource (1-2): 1
Hopefully someone can see what I’m doing wrong here.
Thanks!
-Joe
smime.p7s
Description: S/MIME cryptographic signature
--
[Bacula-users] VolumeRetention, Prune and Pool clarification
Hi all! I’ve been trying to figure something out, and I’m hoping someone can shed some light on this. In the documentation regarding Automatic Volume Recycling, it has this to say about pruning: "When a Job requests another volume and there are no Volumes with Volume Status Append available, Bacula will begin volume pruning. This means that all Jobs that are older than the VolumeRetention period will be pruned from every Volume that has Volume Status Full or Used and has Recycle set to yes.” My question is, does this apply to the pool in question? Or all jobs from all pools that are older than their respective Volume Retention periods pruned? For example, if I have jobs/volumes in Pool A that are older than their volume retention period, but the job that’s running is looking for a volume in Pool B, will it cause the jobs/Volumes in Pool A to get pruned from the database? Thanks! -Joe Rhodes smime.p7s Description: S/MIME cryptographic signature -- Want fast and easy access to all the code in your enterprise? Index and search up to 200,000 lines of code with a free copy of Black Duck Code Sight - the same software that powers the world's largest code search on Ohloh, the Black Duck Open Hub! Try it now. http://p.sf.net/sfu/bds___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] RecycleOldestVolume clarification
Hi again! I’m trying to wrap my head around the option “RecycleOldestVolume=yes”. I’m not entirely clear where it would do anything that isn’t all ready part of the volume recycling algorithm. The documentation explains the volume search/recycling algorithm, including this step: • Prune volumes applying Volume retention period (Volumes with VolStatus Full, Used, or Append are pruned). Note, even if all the File and Job records are pruned from a Volume, the Volume will not be marked Purged until the Volume retention period expires. Then in a later step, it says this: • Prune the oldest Volume if RecycleOldestVolume=yes (the Volume with the oldest LastWritten date and VolStatus equal to Full, Recycle, Purged, Used, or Append is chosen). This record ensures that all retention periods are properly respected. Wouldn’t this all ready have been done by the pruning in the previous step? If RecycleOldestVolume respects all the volume retention periods, I’m not clear how it differs from the normal Prune operation that would otherwise happen. Or am I missing the purpose of this directive? Perhaps a simple scenario where it would make a difference? Thanks! -Joe Rhodes smime.p7s Description: S/MIME cryptographic signature -- Want fast and easy access to all the code in your enterprise? Index and search up to 200,000 lines of code with a free copy of Black Duck Code Sight - the same software that powers the world's largest code search on Ohloh, the Black Duck Open Hub! Try it now. http://p.sf.net/sfu/bds___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Recycling issue
I’m also running into this issue.
In my case, I’ve setup a disk-based auto-changer (vchanger) with two pools:
Full and Weekly. Each pool has multiple drives with 8 volumes on each drive.
The script that mounts the disk also issues an “update slots” command in
bconsole, and I can verify that bacula sees the correct volumes as “InChanger”.
When I swap out disks (weekly), Bacula continues to look for volumes that were
append-able, but on the previous disk (and not currently marked as “InChanger”)
and refuses to write to the append-able volumes on the new disk that are marked
as “InChanger”. It marks all the previous volumes it cannot access as “Error”,
and only then will it begin writing to the currently loaded volumes.
This line from the “Recycling Algorithm” led me to believe that it would use a
volume that was InChanger first: "If the request is for an Autochanger device,
look only for Volumes in the Autochanger (i.e. with InChanger set and that have
the correct Storage device).”
But it seems I’m missing something. I’ve included relevant bits from my
bacula-dir.conf file if anyone can see what I’m doing wrong. I’m getting this
on two systems, one running Bacula 5.2.12 and one running 7.0.4.
My other option is to split these all into separate pools, but that’s going to
make scheduling much harder.
Thanks in advance!
-Joe Rhodes
# Definition of file Virtual Autochanger device
Storage {
Name = removeable-drives
Address = 10.11.1.2
SDPort = 9103
Password = “password"
Device = ExternalDrives
Media Type = File
Maximum Concurrent Jobs = 1
}
# Pools
Pool {
Name = FullSet
Pool Type = Backup
Recycle = yes
AutoPrune = yes
Volume Use Duration = 1 week
#We're on an 2 month rotation
VolumeRetention = 58 days
Maximum Volumes = 32 # 4 drives with 8 volumes on each
Storage = removeable-drives
}
Pool {
Name = Weekly
Pool Type = Backup
Recycle = yes
AutoPrune = yes
Volume Use Duration = 1 week
VolumeRetention = 3 weeks
Maximum Volumes = 40 # 5 drives with 8 volumes on each
Storage = removeable-drives
}
And my bacula-sd.conf:
# Define a Virtual autochanger
Autochanger {
Name = ExternalDrives
Device = ExternalDrives-Drive1
Changer Command = "/usr/local/bin/vchanger -u root %c %o %S %a %d"
Changer Device = "/etc/bacula/vchanger1.conf"
}
Device {
Name = ExternalDrives-Drive1
DriveIndex = 0
Autochanger = yes
DeviceType = File
Media Type = File
Archive Device = /var/lib/bacula/removeable-drives/0/drive0
Random Access = Yes;
AutomaticMount = yes;
RemovableMedia = no;
AlwaysOpen = no;
Maximum Concurrent Jobs = 5;
Maximum Volume Size = 476250 M; # with 8 volumes per drive, this should
leave about
# 4,800 MB free for other things, or about
0.125%
}
> From: Gael Guilmin
> Subject: [Bacula-users] Recycling issue
> To: "bacula-users@lists.sourceforge.net"
>
> Message-ID:
>
> <05520d1d14f2408f8d048f1e1dfe6...@blupr07mb836.namprd07.prod.outlook.com>
>
> Content-Type: text/plain; charset="us-ascii"
>
> Hello,
> I'm facing an issue with Bacula (5.2.12) on a Ubuntu server (12.04.4 LTS): I
> don't know why, but the server, after finishing writing on a tape, tries to
> use a tape that is not in the autochanger, despite the fact that there are
> plenty of online and usable tape in the Scratch pool...
>
> Any idea?
>
> Best regards,
> Gael
> --
> System Administrator / IT
> [PDM_Logo_RGB - registered]
> 78, Avenue du XXeme Corps
> 54000 Nancy
> France
> Phone(Paris): +33 (0)1 49 49 05 25
> Phone(Nancy): +33 (0)3 83 67 66 39
> Cell: +33 (0)6 10 18 48 43
> Web: www.pdgm.com<http://www.pdgm.com/>
--
Infragistics Professional
Build stunning WinForms apps today!
Reboot your WinForms applications with our WinForms controls.
Build a bridge from your legacy apps to the future.
http://pubads.g.doubleclick.net/gampad/clk?id=153845071&iu=/4140/ostg.clktrk___
Bacula-users mailing list
Bacula-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Recycling issue
t; (/var/lib/bacula/removeable-drives/0/drive0). 04-Aug 23:51 mail.roundhouse.local-sd JobId 298: Marking Volume "removeable-drives_0006_0005" in Error in Catalog. 04-Aug 23:51 mail.roundhouse.local-sd JobId 298: Warning: mount.c:212 Open of file device "ExternalDrives-Drive1" (/var/lib/bacula/removeable-drives/0/drive0) Volume "removeable-drives_0006_0005" failed: ERR=file_dev.c:172 Could not open(/var/lib/bacula/removeable-drives/0/drive0,OPEN_READ_WRITE,0640): ERR=No such file or directory 04-Aug 23:51 mail.roundhouse.local-sd JobId 298: Please mount append Volume "removeable-drives_0006_0005" or label a new one for: Job: Server.2014-08-04_18.00.00_33 Storage: "ExternalDrives-Drive1" (/var/lib/bacula/removeable-drives/0/drive0) Pool: Weekly Media type: File 04-Aug 23:56 mail.roundhouse.local-sd JobId 298: Warning: mount.c:212 Open of file device "ExternalDrives-Drive1" (/var/lib/bacula/removeable-drives/0/drive0) Volume "removeable-drives_0006_0006" failed: ERR=file_dev.c:172 Could not open(/var/lib/bacula/removeable-drives/0/drive0,OPEN_READ_WRITE,0640): ERR=No such file or directory 05-Aug 00:01 mail.roundhouse.local-sd JobId 298: Warning: mount.c:212 Open of file device "ExternalDrives-Drive1" (/var/lib/bacula/removeable-drives/0/drive0) Volume "removeable-drives_0006_0006" failed: ERR=file_dev.c:172 Could not open(/var/lib/bacula/removeable-drives/0/drive0,OPEN_READ_WRITE,0640): ERR=No such file or directory 05-Aug 00:05 mail.roundhouse.local-sd JobId 298: Fatal error: Job 298 canceled. 05-Aug 00:05 mail.roundhouse.local-dir JobId 298: Bacula mail.roundhouse.local-dir 7.0.4 (04Jun14): Build OS: x86_64-unknown-linux-gnu ubuntu 14.04 JobId: 298 Job:Server.2014-08-04_18.00.00_33 Backup Level: Incremental, since=2014-08-01 23:18:37 Client: "server" 7.0.4 (04Jun14) x86_64-unknown-linux-gnu,ubuntu,14.04 FileSet:"Linux Server" 2014-07-10 13:39:56 Pool: "Weekly" (From Run Pool override) Catalog:"MyCatalog" (From Client resource) Storage:"removeable-drives" (From Pool resource) Scheduled time: 04-Aug-2014 18:00:00 Start time: 04-Aug-2014 18:00:41 End time: 05-Aug-2014 00:05:51 Elapsed time: 6 hours 5 mins 10 secs Priority: 10 FD Files Written: 0 SD Files Written: 0 FD Bytes Written: 0 (0 B) SD Bytes Written: 0 (0 B) Rate: 0.0 KB/s Software Compression: None VSS:no Encryption: no Accurate: no Volume name(s): Volume Session Id: 110 Volume Session Time:1406833087 Last Volume Bytes: 241 (241 B) Non-fatal FD errors:0 SD Errors: 1 FD termination status: Canceled SD termination status: Canceled Termination:Backup Canceled On Aug 6, 2014, at 1:41 AM, Kern Sibbald wrote: > Hello, > > Probably 95% of all such problems are due to configurations where the > MediaType is not properly set. > > - If you are using any form of autochanger, each autochanger must have a > *unique* MediaType. > > - Each Device configuration in the SD must have a unique MediaType for each > different ArchiveDevice definition. Said the other way around, each > ArchiveDevice definition must have its own unique MediaType. > > The .conf files you list are incomplete, so it is not possible to analyze > them, but when I see: MediaType = File, that raises a red flag that > indicates that there may be other Storage devices with the same MediaType, > and if that is the case, sooner or later, you will find that Bacula is very > confused. > > In the next Bacula version, I will probably modify the SD and possibly the > Dir to require unique MediaTypes, then this problem will be resolved. > > I don't usually supply "support" because I don't have the time, but this > particular problem interests me from the standpoint that I would like to > prevent it, but with no console/job output showing the failures and > incomplete DIR and SD confs I cannot do any more. > > Best regards, > Kern > > On 08/05/2014 03:08 PM, Joe Rhodes wrote: >> I’m also running into this issue. >> >> In my case, I’ve setup a disk-based auto-changer (vchanger) with two pools: >> Full and Weekly. Each pool has multiple drives with 8 volumes on each >> drive. The script that mounts the disk also issues an “update slots” >> command in bconsole, and I can verify that bacula sees the correct volumes >> as “InChanger”. >> >> When
Re: [Bacula-users] Recycling issue
Adding the line:
Autochanger = yes
to my storage definition in my bacula-dir.conf file did indeed solve the issue.
Bacula is now considering volumes only from the ones that are listed as
“InChanger”. Thanks for the pointer.
Cheers!
-Joe
On Aug 6, 2014, at 6:52 AM, Martin Simmons wrote:
>> On Wed, 6 Aug 2014 06:52:06 +, Gael Guilmin said:
>>
>> Here is my Storage definition in bacula-dir.conf:
>> Storage {
>> Name = Autochanger
>> Address = 192.168.27.7 # name or IP address of the server that hosts the
>> daemon storage
>> SDPort = 9103
>> Password = ""
>> Device = "Autochanger" # the same as defined in bacula-sd.conf
>> Media Type = LTO-5
>> }
>
> You need to add
>
> Autochanger = yes
>
> to this bacula-dir definition, otherwise Bacula will not apply the InChanger
> logic. I think this is missing from most of the examples.
>
> __Martin
--
___
Bacula-users mailing list
Bacula-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] wxConsole compile on OS X 10.4
Hello all! I'm trying to get the wxConsole component of bacula to compile on my OS X box, but I'm running into errors: Joe-Rhodes:/Users/joe/Desktop/bacula-1.38.11/src/wx-console root# make /usr/bin/g++ -c -DHAVE_WXCONSOLE -I/usr/lib/wx/include/mac-ansi- release-2.6 -I/usr/include/wx-2.6 -D__WXMAC__ -D_FILE_OFFSET_BITS=64 - D_LARGE_FILES -DNO_GCC_PRAGMA -DHAVE_MACOSX -DNO_GCC_PRAGMA \ -I. -I.. -g -O2 -Wall main.cpp /usr/bin/g++ -c -DHAVE_WXCONSOLE -I/usr/lib/wx/include/mac-ansi- release-2.6 -I/usr/include/wx-2.6 -D__WXMAC__ -D_FILE_OFFSET_BITS=64 - D_LARGE_FILES -DNO_GCC_PRAGMA -DHAVE_MACOSX -DNO_GCC_PRAGMA \ -I. -I.. -g -O2 -Wall console_thread.cpp /usr/include/wx-2.6/wx/list.h: In constructor 'wxListKey::wxListKey (const wxChar*)': /usr/include/wx-2.6/wx/list.h:366: error: 'bad_call_on_strdup_use_bstrdup' was not declared in this scope /usr/include/wx-2.6/wx/list.h: In constructor 'wxListKey::wxListKey (const wxString&)': /usr/include/wx-2.6/wx/list.h:368: error: 'bad_call_on_strdup_use_bstrdup' was not declared in this scope make: *** [console_thread.o] Error 1 Here are the important specs: OS X 10.4.7 (ppc) Bacula 1.38.11 wxWidgets (wxMac) 2.6.3 I've complied wxWidgets with: ./configure --prefix=/usr and bacula claims to have found my wxWidgets 2.6 libraries. (OS X ships with 2.5.) Large file support: no Bacula conio support: yes -ltermcap readline support: no TCP Wrappers support: no TLS support:no ZLIB support: yes enable-smartalloc: yes enable-gnome: no enable-wx-console: yes wxWidgets 2.6 enable-tray-monitor: client-only:no build-dird: yes build-stored: yes ACL support:yes Python support: no I've found one other thread in the mailing list that pertained to this, but I didn't see a resolution. I haven't been able to turn much up on Google either. Any help would be greatly appreciated. Cheers! -Joe Rhodes - Take Surveys. Earn Cash. Influence the Future of IT Join SourceForge.net's Techsay panel and you'll get the chance to share your opinions on IT & business topics through brief surveys -- and earn cash http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
