Re: [Bacula-users] Restore from dead client

Birre,

Thanks for the reply. I guess this is where I get lost...

The fifo is reading a file that was created in the pre-process called mail.tar. The mail.tar is made from the following directories /opt/zimbra and /var/mail/zimbra. This is where the Zimbra files and mailstore were kept.

This pre-process is a script that has this :

MailBackup.bash

#!/bin/bash

exec >/dev/null

MKDIR="/bin/mkdir"
MKFIFO="/usr/bin/mkfifo"
RM="/bin/rm"
TAR="/bin/tar"

DEFCODE=0
DUMPBASE="/data/backups"

errCode=${DEFCODE}
mailDir="/var/mail/zimbra"
zimbraDir="/opt/zimbra"

Main()
{
service zimbra stop

RunMailRestore

service zimbra start

ExitScript ${errCode}
}

RunMailRestore()
{
EXTENSION=".tar"

dumpDir="${DUMPBASE}/mail"
fifoDir="${dumpDir}/fifo"

RebuildFifoDir

${MKFIFO} ${fifoDir}/mail${EXTENSION}

${TAR} -xpf ${fifoDir}/mail${EXTENSION} 2>&1 </dev/null &
}

RebuildFifoDir()
{
if [ -d ${fifoDir} ]
then
${RM} -rf ${fifoDir}
fi

${MKDIR} -p ${fifoDir}
}

ExitScript()
{
exit ${1}
}

Main

The restore script simply does a tar xpf instead of a tar cpf.

Below is the client config for my mail server

#********************************************************************************
# bluewhale
#********************************************************************************
Client {
Name = bluewhale
Address = bluewhale.bnesystems.com
Catalog = BS01-Catalog
Password = "xxxxxxx"
FileRetention = 30 days
JobRetention = 30 days
AutoPrune = yes
MaximumConcurrentJobs = 1
}
Job {
Name = Backup_mail_bluewhale
Enabled = yes
Type = Backup
Level = Full
Accurate = yes
Client = bluewhale
FileSet = Full_mail_bluewhale
Messages = Standard
Pool = Pool_mail_bluewhale
Schedule = Schedule_mail_bluewhale
MaxRunTime = 6 hours
MaxStartDelay = 1 hour
ClientRunBeforeJob = "/nas/users/admin/scripts/MailBackup.bash"
ClientRunAfterJob = "/nas/users/admin/scripts/MailCleanup.bash"
AllowDuplicateJobs = no
CancelQueuedDuplicates = yes
RunScript {
Command="/nas/users/admin/scripts/MailCleanup.bash"
RunsWhen=After
RunsOnFailure=yes
RunsOnClient=yes
RunsOnSuccess=no
}
}
Job {
Name = Restore_mail_bluewhale
FileSet = Full_mail_bluewhale
Type = Restore
Pool = Pool_mail_bluewhale
Client = bluewhale
Messages = Standard
ClientRunBeforeJob = "/nas/users/admin/scripts/MailRestore.bash"
ClientRunAfterJob = "/nas/users/admin/scripts/MailCleanup.bash"
RunScript {
Command="/nas/users/admin/scripts/MailCleanup.bash"
RunsWhen=After
RunsOnFailure=yes
RunsOnClient=yes
RunsOnSuccess=no
}
}
Job {
Name = Backup_os_bluewhale
Enabled = yes
Type = Backup
Level = Full
Accurate = yes
Client = bluewhale
FileSet = Full_os_bluewhale
Messages = Standard
Pool = Pool_os_bluewhale
Schedule = Schedule_os_bluewhale
MaxRunTime = 6 hours
IncrementalMaxRunTime = 2 hours
DifferentialMaxRunTime = 4 hours
AllowDuplicateJobs = no
CancelQueuedDuplicates = yes
}
Job {
Name = Restore_os_bluewhale
FileSet = Full_os_bluewhale
Type = Restore
Pool = Pool_os_bluewhale
Client = bluewhale
Messages = Standard
}
Pool {
Name = Pool_mail_bluewhale
PoolType = Backup
Storage = Storage_bluewhale
MaximumVolumeJobs = 1
CatalogFiles = yes
AutoPrune = yes
VolumeRetention = 365 days
Recycle = yes
LabelFormat = "mail-"
}
Pool {
Name = Pool_os_bluewhale
PoolType = Backup
Storage = Storage_bluewhale
MaximumVolumeJobs = 1
CatalogFiles = yes
AutoPrune = yes
VolumeRetention = 365 days
Recycle = yes
LabelFormat = "os-"
}
Storage {
Name = Storage_bluewhale
Address = 10.10.10.199
SDPort = 9103
Password = "imadirector"
Device = File_bluewhale
MediaType = NAS_bluewhale
MaximumConcurrentJobs = 1
}
Schedule {
Name = Schedule_mail_bluewhale
Run = Level=Full sun-sat at 01:00
}
Schedule {
Name = Schedule_os_bluewhale
Run = Level=Full 1st,3rd sun at 00:00
Run = Level=Differential 2nd,4th sun at 00:00
Run = Level=Incremental mon-sat at 00:00
}
FileSet {
Name = Full_mail_bluewhale
Include {
Options {
signature=SHA1
readfifo=yes
}
File="/data/backups/mail/fifo/mail.tar"
}
}
FileSet {
Name = Full_os_bluewhale
Include {
Options {
compression=GZIP5
_onefs_=no
recurse=yes
hardlinks=yes
WildDir=".gvfs"
WildDir="lost+found"
Exclude=yes
}
File="/"
}
Exclude {
File="/cdrom"
File="/data/backups"
File="/data/downloads"
File="/data/spool"
File="/data/tmp"
File="/dev"
File="/mnt"
File="/nas"
File="/opt/zimbra"
File="/proc"
File="/sys"
File="/tmp"
File="/var/mail/zimbra"
}
}

So why doesn't the name pipe rebuild the file??

Thanks --Kenny

On Fri, Sep 5, 2014 at 2:30 PM, Birger Blixt <birre AT virtualgate DOT org> wrote:

2014-09-05 17:56, Kenny Noe skrev:
>
> I get a zero size mail.tar file in the destination location, so I know
> the changes in location works. Should I post mf conf files?? Please
> advise...
>
> Thanks --kenny
It's not a file, it's a named pipe (fifo)

/nas/users/admin/backups/data/backups/mail/fifo/mail.tar
You must figure out what was reading that fifo and where it stored the data.

/Birre

------------------------------------------------------------------------------
Slashdot TV.  
Video for Nerds.  Stuff that matters.
http://tv.slashdot.org/

_______________________________________________
Bacula-users mailing list
Bacula-users AT lists.sourceforge DOT net
https://lists.sourceforge.net/lists/listinfo/bacula-users