Shulou(Shulou.com)05/31 Report--

This article is about how mkcephfs creates clusters. Xiaobian thinks it is quite practical, so share it with everyone for reference. Let's follow Xiaobian and have a look.

$0 is mkcephfs.

1. Verify the conf file and create a random temporary directory:

dir = /tmp/mkcephfs.jYWWOaxaTa

2 . Execution: $0 --prepare-monmap -d $dir -c $conf

1) Get mon list, mon id, and mon address:

mons=`$CCONF -c $conf -l mon | egrep -v '^mon$' | sort`

id=`echo $name | cut -c 4- | sed 's/^\\.// '`

get_conf addr "" "mon addr"

Then args: args = --add 0 192.168.100.61:6789

2) Create monmap in $dir/monmap

$BINDIR/monmaptool --create --clobber $args --print $monmap || exit 1

In this, fsid is generated and epoch0 is written to/tmp/mkcephfs.jYWWOaxaTa/monmap

3) Copy the ceph.conf file:

cp $conf $dir/conf

At this point, the $0 --prepare-monmap -d $dir -c $conf command ends.

3. Just take osd_list and mds_lilst to form name_list, traverse name_llist and execute:

1) If it is a remote host, then call scp to push conf and monmap to the temporary directory of the remote host $host:$rdir:

rdir=`mktemp -u /tmp/mkfs.ceph.XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX`

2）do_root_cmd "$0 -d $rdir --init-daemon $name"

s1. Check/var/run/ceph/$name.pid and/var/run/ceph/$name.asok and create the appropriate directories

s2 . If the $type of $name is osd, then:

$BINDIR/ceph-osd -c $conf --monmap $dir/monmap -i $id --mkfs --mkkey

get_conf osd_data "/var/lib/ceph/osd/ceph-$id" "osd data"

get_conf osd_keyring "$osd_data/keyring" "keyring"

$BINDIR/ceph-authtool -p -n $name $osd_keyring > $dir/key.$ name

s3. If $type is mds, then:

get_conf mds_data "/var/lib/ceph/mds/ceph-$id" "mds data"

get_conf mds_keyring "$mds_data/keyring" "keyring"

test -d $mds_data || mkdir -p $mds_data

echo "creating private key for $name keyring $mds_keyring"

$BINDIR/ceph-authtool --create-keyring --gen-key -n $name $mds_keyring

$BINDIR/ceph-authtool -p -n $name $mds_keyring > $dir/key.$ name

3)If with--mkfs argument, format partition as well

do_root_cmd "$0 -d $rdir --prepare-osdfs $name"

4) If it is a remote host, collect keys from remote nodes

scp -q $host:$rdir/key.$ name $dir

4. Prepare monitors, execute $0 -d $dir --prepare-mon $moreargs:

At this point, moeargs already contains more parameters: numosd osdmap crushmapsrc crushmap

1) If useosdmap is not empty, use useosdmap: cp $useosdmap $dir/osdmap, otherwise create a generic osdmap from the conf file:

$BINDIR/osdmaptool --create-from-conf $dir/osdmap -c $conf

2)Get crushmapsrc:

get_conf crushmapsrc "" "crush map src" mon global

If $crushmapsrc is not empty, we get a map file from crushmapsrc and save it as $dir/crushmap:

$BINDIR/crushtool -c $crushmapsrc -o $dir/crushmap

3) Get crushmap:

get_conf crushmap "$usecrushmap" "crush map" mon global

If $crushmap is not empty, import $crushmap into $dir/crushmap:

$BINDIR/osdmaptool --import-crush $crushmap $dir/osdmap

4) Generate admin keyring and save it to $dir/keyring.admin:

$BINDIR/ceph-authtool --create-keyring --gen-key -n client.admin $dir/keyring.admin

5) Create the initial monitor keyring:

s1. cp $dir/keyring.admin $dir/keyring.mon

s2. $BINDIR/ceph-authtool -n client.admin --set-uid=0 \

--cap mon 'allow *' \

--cap osd 'allow *' \

--cap mds 'allow' \

$dir/keyring.mon

s3. $BINDIR/ceph-authtool --gen-key -n mon. $dir/keyring.mon --cap mon 'allow *'

6)Traverse $dir/key.*:

If ktype is osd:

$BINDIR/ceph-authtool -n $kname --add-key $secret $dir/keyring.mon \

--cap mon 'allow rwx' \

--cap osd 'allow *'

If type is MDS:

$BINDIR/ceph-authtool -n $kname --add-key $secret $dir/keyring.mon \

--cap mon "allow rwx" \

--cap osd 'allow *' \

--cap mds 'allow'

5. Traverse mon_list and execute:

1)Push Profile

If it is a remote host, push the local $dir/* to $rdir first:

rdir=`mktemp -u /tmp/mkfs.ceph.XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX`

Otherwise cp $dir/conf /etc/ceph/ceph.conf if local host

2)Start the mon daemon:

do_root_cmd "$0 -d $rdir --init-daemon $name"

s1. Check/var/run/ceph/$name.pid and/var/run/ceph/$name.asok and create the appropriate directories

S2. If $type is mon, then:

get_conf mon_data "/var/lib/ceph/mon/ceph-$id" "mon data"

mkdir -p "$mon_data"

$BINDIR/ceph-mon -c $conf --mkfs -i $id --monmap $dir/monmap --osdmap $dir/osdmap -k $dir/keyring.mon

6 . Get adminkerying:

get_conf adminkeyring "/etc/ceph/keyring" "keyring" global

cp $dir/keyring.admin $adminkeyring

Thank you for reading! About "mkcephfs how to create clusters" this article is shared here, I hope the above content can be of some help to everyone, so that everyone can learn more knowledge, if you think the article is good, you can share it to let more people see it!

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.