Zmbkpose
Zmbkpose copied to clipboard
bggo
Restauring from FULL curl: option --data-binary: out of memory
Hi Guys,
I have a small problem :) When i try to restore mailbox from my full backup (~3GB ) i receiving this error:
2014-02-23 21:45:55 DEBUG: [email protected] - Restauring from 23/02/2014 21:08:03 FULL curl: option --data-binary: out of memory curl: try 'curl --help' or 'curl --manual' for more information
enybody has sollution for this?
p.s. when i try to make restore from another user (mailbox 200mb ) its working without any problems
curl cannot get sufficient memory for build the request. Try save memory skipping the web service. Use zimbra user, in your zimbra server, and import the backup as follow : $ /opt/zimbra/bin/zmmailbox -z -m UserEmailAddress postRestURL "//?fmt=tgz&resolve=reset" backup.tgz replace "UserEmailAddress", and "backup.tgz" with the correct values.
You may get the following error: ERROR: zclient.IO_ERROR (Broken pipe) (cause: java.net.SocketException Broken pipe)
In this case, try the follow:
Agree http://www.zimbra.com/forums/administrators/47949-recovery-tgz-mail-file-fails.html
, you can try to import each message using
$ zmmailbox -z -m
Note1: zmmailbox can use a directory argument, but it not examines recursivly, and not filters eml files. Note2: Each directory must exist on the destination mailbox, otherwise you get an error. See below for a script for filter eml files, and create each directory in destination mailbox. untar your backup.tgz file, and use zimbra user for execute the scritp as follow:
$ import_backup.sh [email protected] your_decompress_backup_dir
and patience .....
############ import_backup.sh ####################################
#!/bin/bash
set -e
email=$1
bkdir="$2"
zmmailbox_bin=/opt/zimbra/bin/zmmailbox
if [ ! -x $zmmailbox_bin ] ;then echo "$zmmailbox_bin does not exists" ; exit ; fi
mb_folders=/tmp/mbox_folders
cd "$bkdir"
debug(){ echo "$(date ): $@" ; }
create_folder(){
debug "Creating folder \"$1\" "
$zmmailbox_bin -v -z -m $email createFolder "$1"
}
#Get actual folders
while read Id View Unread Msg_Count Path ;do
[ -z "$Path" ] || echo $Path
done<<EOF |grep '^/.' > $mb_folders
$($zmmailbox_bin -v -z -m $email getAllFolders)
EOF
#Import messages
while read dir;do
#Folder exists?
egrep -q "^${dir#.}$" $mb_folders || create_folder "${dir#.}"
#Import messages in dir
find "$dir" -mindepth 1 -maxdepth 1 -type f -regex '.*\.eml' -print0 |\
xargs -L 1 -0 -r $zmmailbox_bin -v -z -m $email addMessage "${dir#.}"
done<<EOF
$(find . -mindepth 1 -type d )
EOF
Hi, just tried the script of the last comment but there's something wrong with it near line 23-24, could you re-post it?
Sorry, I modified the previous post . This is a display problem because GitHub use Markdown parse for text styles, and ate some characters. Try now please.
Hi, i tried your script but is extremely slow, I checked it and found the problem (and a solution). The message import is done with xargs, but zmmailbox is called for every message:
#Import messages in dir
find "$dir" -mindepth 1 -maxdepth 1 -type f -regex '.*\.eml' -print0 |\
xargs -L 1 -0 -r $zmmailbox_bin -v -z -m $email addMessage "${dir#.}"
I swapped these previous line with:
$zmmailbox_bin -v -z -m $email addMessage "${dir#.}" "$dir"/*eml
In this way all the messages in a folder are imported in a single run of zmmailbox.
Hi, I'm trying the scripts but I get this error for some messages:
Exception in thread "main" java.io.FileNotFoundException: ./Inbox!49/0000220914-CD SF LURIN_ ACTA DE REUNI??N 39.eml (No such file or directory)
at java.base/java.io.FileInputStream.open0(Native Method)
at java.base/java.io.FileInputStream.open(FileInputStream.java:219)
at java.base/java.io.FileInputStream.
The issue of lack of memory was resolved for me by adding physical memory to the server. The server consumed 7 GB of RAM, the archive took up 17 GB. 7+17=24 GB of RAM. I allocated 32 GB of RAM. The recovery went without problems.
