Re: [ClusterLabs] stonith-ng - performing action 'monitor' timed out with signal 15
Hi, some updates about this? Thank you Il Mer 4 Set 2019, 10:46 Marco Marino ha scritto: > First of all, thank you for your support. > Andrey: sure, I can reach machines through IPMI. > Here is a short "log": > > #From ld1 trying to contact ld1 > [root@ld1 ~]# ipmitool -I lanplus -H 192.168.254.250 -U root -P XX > sdr elist all > SEL | 72h | ns | 7.1 | No Reading > Intrusion| 73h | ok | 7.1 | > iDRAC8 | 00h | ok | 7.1 | Dynamic MC @ 20h > ... > > #From ld1 trying to contact ld2 > ipmitool -I lanplus -H 192.168.254.251 -U root -P XX sdr elist all > SEL | 72h | ns | 7.1 | No Reading > Intrusion| 73h | ok | 7.1 | > iDRAC7 | 00h | ok | 7.1 | Dynamic MC @ 20h > ... > > > #From ld2 trying to contact ld1: > root@ld2 ~]# ipmitool -I lanplus -H 192.168.254.250 -U root -P X sdr > elist all > SEL | 72h | ns | 7.1 | No Reading > Intrusion| 73h | ok | 7.1 | > iDRAC8 | 00h | ok | 7.1 | Dynamic MC @ 20h > System Board | 00h | ns | 7.1 | Logical FRU @00h > . > > #From ld2 trying to contact ld2 > [root@ld2 ~]# ipmitool -I lanplus -H 192.168.254.251 -U root -P sdr > elist all > SEL | 72h | ns | 7.1 | No Reading > Intrusion| 73h | ok | 7.1 | > iDRAC7 | 00h | ok | 7.1 | Dynamic MC @ 20h > System Board | 00h | ns | 7.1 | Logical FRU @00h > > > Jan: Actually the cluster uses /etc/hosts in order to resolve names: > 172.16.77.10ld1.mydomain.it ld1 > 172.16.77.11ld2.mydomain.it ld2 > > Furthermore I'm using ip addresses for ipmi interfaces in the > configuration: > [root@ld1 ~]# pcs stonith show fence-node1 > Resource: fence-node1 (class=stonith type=fence_ipmilan) > Attributes: ipaddr=192.168.254.250 lanplus=1 login=root passwd=X > pcmk_host_check=static-list pcmk_host_list=ld1.mydomain.it > Operations: monitor interval=60s (fence-node1-monitor-interval-60s) > > > Any idea? > How can I reset the state of the cluster without downtime? "pcs resource > cleanup" is enough? > Thank you, > Marco > > > Il giorno mer 4 set 2019 alle ore 10:29 Jan Pokorný > ha scritto: > >> On 03/09/19 20:15 +0300, Andrei Borzenkov wrote: >> > 03.09.2019 11:09, Marco Marino пишет: >> >> Hi, I have a problem with fencing on a two node cluster. It seems that >> >> randomly the cluster cannot complete monitor operation for fence >> devices. >> >> In log I see: >> >> crmd[8206]: error: Result of monitor operation for fence-node2 on >> >> ld2.mydomain.it: Timed Out >> > >> > Can you actually access IP addresses of your IPMI ports? >> >> [ >> Tangentially, interesting aspect beyond that and applicable for any >> non-IP cross-host referential needs, which I haven't seen mentioned >> anywhere so far, is the risk of DNS resolution (when /etc/hosts will >> come short) getting to troubles (stale records, port blocked, DNS >> server overload [DNSSEC, etc.], IPv4/IPv6 parallel records that the SW >> cannot handle gracefully, etc.). In any case, just a single DNS >> server would apparently be an undesired SPOF, and would be unfortunate >> when unable to fence a node because of that. >> >> I think the most robust approach is to use IP addresses whenever >> possible, and unambiguous records in /etc/hosts when practical. >> ] >> >> >> As attachment there is >> >> - /var/log/messages for node1 (only the important part) >> >> - /var/log/messages for node2 (only the important part) <-- Problem >> starts >> >> here >> >> - pcs status >> >> - pcs stonith show (for both fence devices) >> >> >> >> I think it could be a timeout problem, so how can I see timeout value >> for >> >> monitor operation in stonith devices? >> >> Please, someone can help me with this problem? >> >> Furthermore, how can I fix the state of fence devices without downtime? >> >> -- >> Jan (Poki) >> ___ >> Manage your subscription: >> https://lists.clusterlabs.org/mailman/listinfo/users >> >> ClusterLabs home: https://www.clusterlabs.org/ > > ___ Manage your subscription: https://lists.clusterlabs.org/mailman/listinfo/users ClusterLabs home: https://www.clusterlabs.org/
Re: [ClusterLabs] stonith-ng - performing action 'monitor' timed out with signal 15
First of all, thank you for your support. Andrey: sure, I can reach machines through IPMI. Here is a short "log": #From ld1 trying to contact ld1 [root@ld1 ~]# ipmitool -I lanplus -H 192.168.254.250 -U root -P XX sdr elist all SEL | 72h | ns | 7.1 | No Reading Intrusion| 73h | ok | 7.1 | iDRAC8 | 00h | ok | 7.1 | Dynamic MC @ 20h ... #From ld1 trying to contact ld2 ipmitool -I lanplus -H 192.168.254.251 -U root -P XX sdr elist all SEL | 72h | ns | 7.1 | No Reading Intrusion| 73h | ok | 7.1 | iDRAC7 | 00h | ok | 7.1 | Dynamic MC @ 20h ... #From ld2 trying to contact ld1: root@ld2 ~]# ipmitool -I lanplus -H 192.168.254.250 -U root -P X sdr elist all SEL | 72h | ns | 7.1 | No Reading Intrusion| 73h | ok | 7.1 | iDRAC8 | 00h | ok | 7.1 | Dynamic MC @ 20h System Board | 00h | ns | 7.1 | Logical FRU @00h . #From ld2 trying to contact ld2 [root@ld2 ~]# ipmitool -I lanplus -H 192.168.254.251 -U root -P sdr elist all SEL | 72h | ns | 7.1 | No Reading Intrusion| 73h | ok | 7.1 | iDRAC7 | 00h | ok | 7.1 | Dynamic MC @ 20h System Board | 00h | ns | 7.1 | Logical FRU @00h Jan: Actually the cluster uses /etc/hosts in order to resolve names: 172.16.77.10ld1.mydomain.it ld1 172.16.77.11ld2.mydomain.it ld2 Furthermore I'm using ip addresses for ipmi interfaces in the configuration: [root@ld1 ~]# pcs stonith show fence-node1 Resource: fence-node1 (class=stonith type=fence_ipmilan) Attributes: ipaddr=192.168.254.250 lanplus=1 login=root passwd=X pcmk_host_check=static-list pcmk_host_list=ld1.mydomain.it Operations: monitor interval=60s (fence-node1-monitor-interval-60s) Any idea? How can I reset the state of the cluster without downtime? "pcs resource cleanup" is enough? Thank you, Marco Il giorno mer 4 set 2019 alle ore 10:29 Jan Pokorný ha scritto: > On 03/09/19 20:15 +0300, Andrei Borzenkov wrote: > > 03.09.2019 11:09, Marco Marino пишет: > >> Hi, I have a problem with fencing on a two node cluster. It seems that > >> randomly the cluster cannot complete monitor operation for fence > devices. > >> In log I see: > >> crmd[8206]: error: Result of monitor operation for fence-node2 on > >> ld2.mydomain.it: Timed Out > > > > Can you actually access IP addresses of your IPMI ports? > > [ > Tangentially, interesting aspect beyond that and applicable for any > non-IP cross-host referential needs, which I haven't seen mentioned > anywhere so far, is the risk of DNS resolution (when /etc/hosts will > come short) getting to troubles (stale records, port blocked, DNS > server overload [DNSSEC, etc.], IPv4/IPv6 parallel records that the SW > cannot handle gracefully, etc.). In any case, just a single DNS > server would apparently be an undesired SPOF, and would be unfortunate > when unable to fence a node because of that. > > I think the most robust approach is to use IP addresses whenever > possible, and unambiguous records in /etc/hosts when practical. > ] > > >> As attachment there is > >> - /var/log/messages for node1 (only the important part) > >> - /var/log/messages for node2 (only the important part) <-- Problem > starts > >> here > >> - pcs status > >> - pcs stonith show (for both fence devices) > >> > >> I think it could be a timeout problem, so how can I see timeout value > for > >> monitor operation in stonith devices? > >> Please, someone can help me with this problem? > >> Furthermore, how can I fix the state of fence devices without downtime? > > -- > Jan (Poki) > ___ > Manage your subscription: > https://lists.clusterlabs.org/mailman/listinfo/users > > ClusterLabs home: https://www.clusterlabs.org/ ___ Manage your subscription: https://lists.clusterlabs.org/mailman/listinfo/users ClusterLabs home: https://www.clusterlabs.org/
[ClusterLabs] stonith-ng - performing action 'monitor' timed out with signal 15
Hi, I have a problem with fencing on a two node cluster. It seems that randomly the cluster cannot complete monitor operation for fence devices. In log I see: crmd[8206]: error: Result of monitor operation for fence-node2 on ld2.mydomain.it: Timed Out As attachment there is - /var/log/messages for node1 (only the important part) - /var/log/messages for node2 (only the important part) <-- Problem starts here - pcs status - pcs stonith show (for both fence devices) I think it could be a timeout problem, so how can I see timeout value for monitor operation in stonith devices? Please, someone can help me with this problem? Furthermore, how can I fix the state of fence devices without downtime? Thank you PCS STATUS root@ld1 ~]# pcs status Cluster name: ldcluster Stack: corosync Current DC: ld1.mydomain.it (version 1.1.19-8.el7_6.4-c3c624ea3d) - partition with quorum Last updated: Tue Sep 3 09:37:27 2019 Last change: Thu Jul 4 21:36:07 2019 by root via cibadmin on ld1.mydomain.it 2 nodes configured 10 resources configured Online: [ ld1.mydomain.it ld2.mydomain.it ] Full list of resources: fence-node1(stonith:fence_ipmilan):Stopped fence-node2(stonith:fence_ipmilan):Stopped Master/Slave Set: DrbdResClone [DrbdRes] Masters: [ ld1.mydomain.it ] Slaves: [ ld2.mydomain.it ] HALVM (ocf::heartbeat:LVM): Started ld1.mydomain.it PgsqlFs(ocf::heartbeat:Filesystem):Started ld1.mydomain.it PostgresqlD(systemd:postgresql-9.6.service): Started ld1.mydomain.it LegaldocapiD (systemd:legaldocapi.service): Started ld1.mydomain.it PublicVIP (ocf::heartbeat:IPaddr2): Started ld1.mydomain.it DefaultRoute (ocf::heartbeat:Route): Started ld1.mydomain.it Failed Actions: * fence-node1_start_0 on ld1.mydomain.it 'unknown error' (1): call=221, status=Timed Out, exitreason='', last-rc-change='Wed Aug 21 12:49:00 2019', queued=0ms, exec=20006ms * fence-node2_start_0 on ld1.mydomain.it 'unknown error' (1): call=222, status=Timed Out, exitreason='', last-rc-change='Wed Aug 21 12:49:00 2019', queued=1ms, exec=20013ms * fence-node1_start_0 on ld2.mydomain.it 'unknown error' (1): call=182, status=Timed Out, exitreason='', last-rc-change='Wed Aug 21 14:26:09 2019', queued=0ms, exec=20006ms * fence-node2_start_0 on ld2.mydomain.it 'unknown error' (1): call=176, status=Timed Out, exitreason='', last-rc-change='Wed Aug 21 12:48:40 2019', queued=1ms, exec=20008ms Daemon Status: corosync: active/disabled pacemaker: active/disabled pcsd: active/enabled [root@ld1 ~]# STONITH SHOW### [root@ld1 ~]# pcs stonith show fence-node1 Resource: fence-node1 (class=stonith type=fence_ipmilan) Attributes: ipaddr=192.168.254.250 lanplus=1 login=root passwd=XXX pcmk_host_check=static-list pcmk_host_list=ld1.mydomain.it Operations: monitor interval=60s (fence-node1-monitor-interval-60s) [root@ld1 ~]# pcs stonith show fence-node2 Resource: fence-node2 (class=stonith type=fence_ipmilan) Attributes: ipaddr=192.168.254.251 lanplus=1 login=root passwd= pcmk_host_check=static-list pcmk_host_list=ld2.mydomain.it delay=12 Operations: monitor interval=60s (fence-node2-monitor-interval-60s) [root@ld1 ~]# ###NODE 2 /var/log/messages## Aug 21 12:48:40 ld2 stonith-ng[8202]: notice: Child process 46006 performing action 'monitor' timed out with signal 15 Aug 21 12:48:40 ld2 stonith-ng[8202]: notice: Operation 'monitor' [46006] for device 'fence-node2' returned: -62 (Timer expired) Aug 21 12:48:40 ld2 crmd[8206]: error: Result of monitor operation for fence-node2 on ld2.mydomain.it: Timed Out Aug 21 12:48:40 ld2 stonith-ng[8202]: notice: On loss of CCM Quorum: Ignore Aug 21 12:48:40 ld2 crmd[8206]: notice: Result of stop operation for fence-node2 on ld2.mydomain.it: 0 (ok) Aug 21 12:48:40 ld2 stonith-ng[8202]: notice: On loss of CCM Quorum: Ignore Aug 21 12:48:40 ld2 stonith-ng[8202]: notice: On loss of CCM Quorum: Ignore Aug 21 12:48:40 ld2 stonith-ng[8202]: notice: On loss of CCM Quorum: Ignore Aug 21 12:48:59 ld2 stonith-ng[8202]: notice: On loss of CCM Quorum: Ignore Aug 21 12:49:00 ld2 stonith-ng[8202]: notice: Child process 46053 performing action 'monitor' timed out with signal 15 Aug 21 12:49:00 ld2 stonith-ng[8202]: notice: Operation 'monitor' [46053] for device 'fence-node2' returned: -62 (Timer expired) Aug 21 12:49:00 ld2 crmd[8206]: error: Result of start operation for fence-node2 on ld2.mydomain.it: Timed Out Aug 21 12:49:00 ld2 stonith-ng[8202]: notice: On loss of CCM Quorum: Ignore Aug 21 12:49:00 ld2 stonith-ng[8202]: notice: On loss of CCM Quorum: Ignore Aug 21 12:49:00 ld2 stonith-ng[8202]: notice: On loss of CCM Quorum: Ignore Aug 21 12:49:00 ld2 stonith-ng[8202]: notice: On loss of CCM Quorum: Ignore Aug 21 12:49:00 ld2 crmd[8206]:
Re: [ClusterLabs] HALVM monitor action fail on slave node. Possible bug?
Hi Emmanuel, thank you for you support. I did a lot of checks during the WE
and there are some updates:
- Main problem is that ocf:heartbeat:LVM is old. The current version on
centos 7 is 3.9.5 (package resource-agents). More precisely, in 3.9.5 the
monitor function has one important assumption: the underlying storage is
shared between all nodes in the cluster. So the monitor function checks the
presence of the volume group on all nodes. From version 3.9.6 this is not
the normal behavior and the monitor function (LVM_status) returns
$OCF_NOT_RUNNING from slaves nodes without errors. You can check this in
the file /usr/lib/ocf/resource.d/heartbeat/LVM in lines 340-351 that
disappears in version 3.9.6.
Obviously this is not error, but an important change in the cluster
architecture because I need to use drbd in dual primary mode when version
3.9.5 is used. My personal idea is that drbd in dual primary mode with lvm
is not a good idea due to the fact that I don't need an active/active
cluster.
Anyway, thank you for your time again
Marco
2018-04-13 15:54 GMT+02:00 emmanuel segura <emi2f...@gmail.com>:
> the first thing that you need to configure is the stonith, because you
> have this constraint "constraint order promote DrbdResClone then start
> HALVM"
>
> To recover and promote drbd to master when you crash a node, configurare
> the drbd fencing handler.
>
> pacemaker execute monitor in both nodes, so this is normal, to test why
> monitor fail, use ocf-tester
>
> 2018-04-13 15:29 GMT+02:00 Marco Marino <marino@gmail.com>:
>
>> Hello, I'm trying to configure a simple 2 node cluster with drbd and
>> HALVM (ocf:heartbeat:LVM) but I have a problem that I'm not able to solve,
>> to I decided to write this long post. I need to really understand what I'm
>> doing and where I'm doing wrong.
>> More precisely, I'm configuring a pacemaker cluster with 2 nodes and only
>> one drbd resource. Here all operations:
>>
>> - System configuration
>> hostnamectl set-hostname pcmk[12]
>> yum update -y
>> yum install vim wget git -y
>> vim /etc/sysconfig/selinux -> permissive mode
>> systemctl disable firewalld
>> reboot
>>
>> - Network configuration
>> [pcmk1]
>> nmcli connection modify corosync ipv4.method manual ipv4.addresses
>> 192.168.198.201/24 ipv6.method ignore connection.autoconnect yes
>> nmcli connection modify replication ipv4.method manual ipv4.addresses
>> 192.168.199.201/24 ipv6.method ignore connection.autoconnect yes
>> [pcmk2]
>> nmcli connection modify corosync ipv4.method manual ipv4.addresses
>> 192.168.198.202/24 ipv6.method ignore connection.autoconnect yes
>> nmcli connection modify replication ipv4.method manual ipv4.addresses
>> 192.168.199.202/24 ipv6.method ignore connection.autoconnect yes
>>
>> ssh-keyget -t rsa
>> ssh-copy-id root@pcmk[12]
>> scp /etc/hosts root@pcmk2:/etc/hosts
>>
>> - Drbd Repo configuration and drbd installation
>> rpm --import https://www.elrepo.org/RPM-GPG-KEY-elrepo.org
>> rpm -Uvh http://www.elrepo.org/elrepo-release-7.0-3.el7.elrepo.noarch
>> .rpm
>> yum update -y
>> yum install drbd84-utils kmod-drbd84 -y
>>
>> - Drbd Configuration:
>> Creating a new partition on top of /dev/vdb -> /dev/vdb1 of type
>> "Linux" (83)
>> [/etc/drbd.d/global_common.conf]
>> usage-count no;
>> [/etc/drbd.d/myres.res]
>> resource myres {
>> on pcmk1 {
>> device /dev/drbd0;
>> disk /dev/vdb1;
>> address 192.168.199.201:7789;
>> meta-disk internal;
>> }
>> on pcmk2 {
>> device /dev/drbd0;
>> disk /dev/vdb1;
>> address 192.168.199.202:7789;
>> meta-disk internal;
>> }
>> }
>>
>> scp /etc/drbd.d/myres.res root@pcmk2:/etc/drbd.d/myres.res
>> systemctl start drbd <-- only for test. The service is disabled at
>> boot!
>> drbdadm create-md myres
>> drbdadm up myres
>> drbdadm primary --force myres
>>
>> - LVM Configuration
>> [root@pcmk1 ~]# lsblk
>> NAMEMAJ:MIN RM SIZE RO TYPE MOUNTPOINT
>> sr0 11:01 1024M 0 rom
>> vda 252:00 20G 0 disk
>> ├─vda1 252:101G 0 part /boot
>> └─vda2 252:20 19G 0 part
>> ├─cl-root 253:00 17G 0 lvm /
>> └─cl-swap 253:102G 0 lvm [SWAP]
>>
[ClusterLabs] HALVM monitor action fail on slave node. Possible bug?
cluster cib lvm_cfg pcs -f lvm_cfg resource create HALVM ocf:heartbeat:LVM volgrpname=havolumegroup pcs -f lvm_cfg constraint colocation add HALVM with master DrbdResClone INFINITY pcs -f lvm_cfg constraint order promote DrbdResClone then start HALVM [root@pcmk1 ~]# pcs -f lvm_cfg constraint Location Constraints: Ordering Constraints: promote DrbdResClone then start HALVM (kind:Mandatory) Colocation Constraints: HALVM with DrbdResClone (score:INFINITY) (rsc-role:Started) (with-rsc-role:Master) Ticket Constraints: [root@pcmk1 ~]# [root@pcmk1 ~]# pcs status Cluster name: ha_cluster Stack: corosync Current DC: pcmk2 (version 1.1.16-12.el7_4.8-94ff4df) - partition with quorum Last updated: Fri Apr 13 15:12:49 2018 Last change: Fri Apr 13 15:05:18 2018 by root via cibadmin on pcmk1 2 nodes configured 2 resources configured Online: [ pcmk1 pcmk2 ] Full list of resources: Master/Slave Set: DrbdResClone [DrbdRes] Masters: [ pcmk2 ] Slaves: [ pcmk1 ] Daemon Status: corosync: active/enabled pacemaker: active/enabled pcsd: active/enabled #[PUSHING NEW CONFIGURATION]# [root@pcmk1 ~]# pcs cluster cib-push lvm_cfg CIB updated [root@pcmk1 ~]# pcs status Cluster name: ha_cluster Stack: corosync Current DC: pcmk2 (version 1.1.16-12.el7_4.8-94ff4df) - partition with quorum Last updated: Fri Apr 13 15:12:57 2018 Last change: Fri Apr 13 15:12:55 2018 by root via cibadmin on pcmk1 2 nodes configured 3 resources configured Online: [ pcmk1 pcmk2 ] Full list of resources: Master/Slave Set: DrbdResClone [DrbdRes] Masters: [ pcmk2 ] Slaves: [ pcmk1 ] HALVM(ocf::heartbeat:LVM):Started pcmk2 Failed Actions: * HALVM_monitor_0 on pcmk1 'unknown error' (1): call=13, status=complete, exitreason='LVM Volume havolumegroup is not available', last-rc-change='Fri Apr 13 15:12:56 2018', queued=0ms, exec=52ms Daemon Status: corosync: active/enabled pacemaker: active/enabled pcsd: active/enabled [root@pcmk1 ~]# ##[TRYING TO CLEANUP RESOURCE CONFIGURATION]## [root@pcmk1 ~]# pcs resource cleanup Waiting for 1 replies from the CRMd. OK [root@pcmk1 ~]# pcs status Cluster name: ha_cluster Stack: corosync Current DC: pcmk2 (version 1.1.16-12.el7_4.8-94ff4df) - partition with quorum Last updated: Fri Apr 13 15:13:18 2018 Last change: Fri Apr 13 15:12:55 2018 by root via cibadmin on pcmk1 2 nodes configured 3 resources configured Online: [ pcmk1 pcmk2 ] Full list of resources: Master/Slave Set: DrbdResClone [DrbdRes] Masters: [ pcmk2 ] Slaves: [ pcmk1 ] HALVM(ocf::heartbeat:LVM):Started pcmk2 Failed Actions: * HALVM_monitor_0 on pcmk1 'unknown error' (1): call=26, status=complete, exitreason='LVM Volume havolumegroup is not available', last-rc-change='Fri Apr 13 15:13:17 2018', queued=0ms, exec=113ms Daemon Status: corosync: active/enabled pacemaker: active/enabled pcsd: active/enabled [root@pcmk1 ~]# # some details about packages and versions: [root@pcmk1 ~]# rpm -qa | grep pacem pacemaker-cluster-libs-1.1.16-12.el7_4.8.x86_64 pacemaker-libs-1.1.16-12.el7_4.8.x86_64 pacemaker-1.1.16-12.el7_4.8.x86_64 pacemaker-cli-1.1.16-12.el7_4.8.x86_64 [root@pcmk1 ~]# rpm -qa | grep coro corosynclib-2.4.0-9.el7_4.2.x86_64 corosync-2.4.0-9.el7_4.2.x86_64 [root@pcmk1 ~]# rpm -qa | grep drbd drbd84-utils-9.1.0-1.el7.elrepo.x86_64 kmod-drbd84-8.4.10-1_2.el7_4.elrepo.x86_64 [root@pcmk1 ~]# cat /etc/redhat-release CentOS Linux release 7.4.1708 (Core) [root@pcmk1 ~]# uname -r 3.10.0-693.21.1.el7.x86_64 [root@pcmk1 ~]# ## So it seems to me that the problem is that the "monitor" action of the ocf:heartbeat:LVM resource is executed on both nodes even if I configured specific colocation and ordering constraints. I don't know where the problem is, but please I need to understand how to solve the issue. Please, if possible I invite someone to reproduce the configuration and possibly the issue. It seems a bug but obviously I'm not sure. What I'm worried is that it should be pacemaker that states where and when one resource should start so probably there is something wrong in my constraints configuration. I'm sorry for this long post. Thank you, Marco ___ Users mailing list: Users@clusterlabs.org https://lists.clusterlabs.org/mailman/listinfo/users Project Home: http://www.clusterlabs.org Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf Bugs: http://bugs.clusterlabs.org
Re: [ClusterLabs] HALVM problem with 2 nodes cluster
Ferenc, regarding the flag use_lvmetad in /usr/lib/ocf/resource.d/heartbeat/LVM I read: "#lvmetad is a daemon that caches lvm metadata to improve the # performance of LVM commands. This daemon should never be used when # volume groups exist that are being managed by the cluster. The lvmetad # daemon introduces a response lag, where certain LVM commands look like # they have completed (like vg activation) when in fact the command # is still in progress by the lvmetad. This can cause reliability issues # when managing volume groups in the cluster. For Example, if you have a # volume group that is a dependency for another application, it is possible # the cluster will think the volume group is activated and attempt to start # the application before volume group is really accesible... lvmetad is bad." in the function LVM_validate_all() Anyway, it's only a warning but there is a good reason. I'm not an expert, I'm studying for a certification and I have a lot of doubts. Thank you for your help Marco 2017-01-18 11:03 GMT+01:00 Ferenc Wágner <wf...@niif.hu>: > Marco Marino <marino@gmail.com> writes: > > > I agree with you for > > use_lvmetad = 0 (setting it = 1 in a clustered environment is an error) > > Where does this information come from? AFAIK, if locking_type=3 (LVM > uses internal clustered locking, that is, clvmd), lvmetad is not used > anyway, even if it's running. So it's best to disable it to avoid > warning messages all around. This is the case with active/active > clustering in LVM itself, in which Pacemaker isn't involved. > > On the other hand, if you use Pacemaker to do active/passive clustering > by appropriately activating/deactivating your VG, this isn't clustering > from the LVM point of view, you don't set the clustered flag on your VG, > don't run clvmd and use locking_type=1. Lvmetad should be perfectly > fine with this in principle (unless it caches metadata of inactive VGs, > which would be stupid, but I never tested this). > > > but I think I have to set > > locking_type = 3 only if I use clvm > > Right. > -- > Feri > ___ Users mailing list: Users@clusterlabs.org http://lists.clusterlabs.org/mailman/listinfo/users Project Home: http://www.clusterlabs.org Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf Bugs: http://bugs.clusterlabs.org
Re: [ClusterLabs] HALVM problem with 2 nodes cluster
Hi Bliu, thank you. I agree with you for use_lvmetad = 0 (setting it = 1 in a clustered environment is an error) but I think I have to set locking_type = 3 only if I use clvm In my case, I'm trying to use LVM so I think that locking_type = 1 is ok. What do you think about? Furthermore, I have an application (managed as a resource in the cluster) that continously create and remove logical volumes in the cluster. Is this a problem? The application uses a custom lvm.conf configuration file where I have volume_list = [ "@pacemaker" ] Thank you 2017-01-18 10:12 GMT+01:00 bliu <b...@suse.com>: > Hi, Marco > > On 01/18/2017 04:45 PM, Marco Marino wrote: > > Hi, I'm trying to realize a cluster with 2 nodes that manages a volume > group. > Basically I have a san connected to both nodes that exposes 1 lun. So both > nodes have a disk /dev/sdb. From one node I did: > fdisk /dev/sdb <- Create a partition with type = 8e (LVM) > pvcreate /dev/sdb1 > vgcreate myvg > > then > > pcs resource create halvm LVM volgrpname=myvg exclusive=true > > Last command fails with an error: "LVM: myvg did not activate correctly" > > Reading /usr/lib/ocf/resource.d/heartbeat/LVM, this happens because it > seems that I need at least one logical volume inside the volume group > before create the resource. Is this correct? > > Yes, you need to create pv, vg before you use cluster to manager it. > > Furthermore, how can I set volume_list in lvm.conf? Actually in lvm.conf I > have: > > Normally, clvm is used in cluster with shared storage as: > locking_type = 3 > use_lvmetad = 0 > > locking_type = 1 > use_lvmetad = 1 > volume_list = [ "vg-with-root-lv" ] > > > Thank you > > > > > ___ > Users mailing list: > Users@clusterlabs.orghttp://lists.clusterlabs.org/mailman/listinfo/users > > Project Home: http://www.clusterlabs.org > Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf > Bugs: http://bugs.clusterlabs.org > > > > ___ > Users mailing list: Users@clusterlabs.org > http://lists.clusterlabs.org/mailman/listinfo/users > > Project Home: http://www.clusterlabs.org > Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf > Bugs: http://bugs.clusterlabs.org > > ___ Users mailing list: Users@clusterlabs.org http://lists.clusterlabs.org/mailman/listinfo/users Project Home: http://www.clusterlabs.org Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf Bugs: http://bugs.clusterlabs.org
[ClusterLabs] HALVM problem with 2 nodes cluster
Hi, I'm trying to realize a cluster with 2 nodes that manages a volume group. Basically I have a san connected to both nodes that exposes 1 lun. So both nodes have a disk /dev/sdb. From one node I did: fdisk /dev/sdb <- Create a partition with type = 8e (LVM) pvcreate /dev/sdb1 vgcreate myvg then pcs resource create halvm LVM volgrpname=myvg exclusive=true Last command fails with an error: "LVM: myvg did not activate correctly" Reading /usr/lib/ocf/resource.d/heartbeat/LVM, this happens because it seems that I need at least one logical volume inside the volume group before create the resource. Is this correct? Furthermore, how can I set volume_list in lvm.conf? Actually in lvm.conf I have: locking_type = 1 use_lvmetad = 1 volume_list = [ "vg-with-root-lv" ] Thank you ___ Users mailing list: Users@clusterlabs.org http://lists.clusterlabs.org/mailman/listinfo/users Project Home: http://www.clusterlabs.org Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf Bugs: http://bugs.clusterlabs.org
Re: [ClusterLabs] Cluster reboot fro maintenance
Maintenance worked perfectly: - cluster in maintenance: crm configure property maintenance-mode=true - update vm os etc - stop corosync/pacemaker - reboot - start corosync/pacemaker - cluster out of maintenance: crm configure property maintenance-mode=false - all resources went up ok Best regards Marco On Mon, 20 Jun 2016 15:42:11 -0500 Ken Gaillot <kgail...@redhat.com> wrote: > On 06/20/2016 07:45 AM, ma...@nucleus.it wrote: > > Hi, > > i have a two node cluster with some vms (pacemaker resources) > > running on the two hypervisors: > > pacemaker-1.0.10 > > corosync-1.3.0 > > > > I need to do maintenance stuff , so i need to: > > - put on maintenance the cluster so the cluster doesn't > > touch/start/stop/monitor the vms > > - update the vms > > - stop the vms > > - stop cluster stuff (corosync/pacemaker) so it do not > > start/stop/monitor vms > > - reboot the hypervisors. > > - start cluster stuff > > - remove maintenance from the cluster stuff so it start all the vms > > > > What is the corret way to do that ( corosync/pacemaker) side ? > > > > > > Best regards > > Marco > > Maintenance mode provides this ability. Set the maintenance-mode > cluster proprerty to true, do whatever you want, then set it back to > false when done. > > That said, I've never used pacemaker/corosync versions that old, so > I'm not 100% sure that applies to those versions, though I would > guess it does. > > ___ > Users mailing list: Users@clusterlabs.org > http://clusterlabs.org/mailman/listinfo/users > > Project Home: http://www.clusterlabs.org > Getting started: > http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf Bugs: > http://bugs.clusterlabs.org ___ Users mailing list: Users@clusterlabs.org http://clusterlabs.org/mailman/listinfo/users Project Home: http://www.clusterlabs.org Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf Bugs: http://bugs.clusterlabs.org
[ClusterLabs] Cluster reboot fro maintenance
Hi, i have a two node cluster with some vms (pacemaker resources) running on the two hypervisors: pacemaker-1.0.10 corosync-1.3.0 I need to do maintenance stuff , so i need to: - put on maintenance the cluster so the cluster doesn't touch/start/stop/monitor the vms - update the vms - stop the vms - stop cluster stuff (corosync/pacemaker) so it do not start/stop/monitor vms - reboot the hypervisors. - start cluster stuff - remove maintenance from the cluster stuff so it start all the vms What is the corret way to do that ( corosync/pacemaker) side ? Best regards Marco ___ Users mailing list: Users@clusterlabs.org http://clusterlabs.org/mailman/listinfo/users Project Home: http://www.clusterlabs.org Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf Bugs: http://bugs.clusterlabs.org
Re: [ClusterLabs] unable to start fence_scsi
Hi Ken, by the way I’ve just also tried with pacemaker 1.1.14 (I builded it from sources into a new RPM) but it doesn’t work > On 18 May 2016, at 11:29, Marco A. Carcano <marco.carc...@itc4u.ch> wrote: > > Hi Ken, > > thank you for the reply > > I tried as you suggested, and now the stonith devices tries to start but > fails. > > I tried this > > pcs stonith create scsi fence_scsi pcmk_host_list="apache-up001.ring0 > apache-up002.ring0 apache-up003.ring0" > pcmk_host_map="apache-up001.ring1=apache-up001.ring0; > apache-up002.ring1=apache-up002.ring0; apache-up003.ring1=apache-up003.ring0" > pcmk_reboot_action="off" > devices="/dev/mapper/36001405973e201b3fdb4a999175b942f" meta > provides="unfencing" op monitor interval=60s > > and even this, adding pcmk_monitor_action="metadata” as suggested in a post > on RH knowledge base (even if the error was quite different) > > pcs stonith create scsi fence_scsi pcmk_host_list="apache-up001.ring0 > apache-up002.ring0 apache-up003.ring0" > pcmk_host_map="apache-up001.ring1=apache-up001.ring0; > apache-up002.ring1=apache-up002.ring0; apache-up003.ring1=apache-up003.ring0" > pcmk_reboot_action="off" > devices="/dev/mapper/36001405973e201b3fdb4a999175b942f" meta > provides="unfencing" pcmk_monitor_action="metadata" op monitor interval=60s > > I’m using CentOS 7.2, pacemaker-1.1.13-10 resource-agents-3.9.5-54 and > fence-agents-scsi-4.0.11-27 > > the error message are Couldn't find anyone to fence (on) apache-up003.ring0 > with any device anderror: Operation on of apache-up003.ring0 by > for crmd.15918@apache-up001.ring0.0599387e: No such device > > Thanks > > Marco > > > May 18 10:37:03 apache-up001 crmd[15918]: notice: State transition S_IDLE -> > S_POLICY_ENGINE [ input=I_PE_CALC cause=C_FSA_INTERNAL > origin=abort_transition_graph ] > May 18 10:37:03 apache-up001 pengine[15917]: notice: On loss of CCM Quorum: > Ignore > May 18 10:37:03 apache-up001 pengine[15917]: notice: Unfencing > apache-up001.ring0: node discovery > May 18 10:37:03 apache-up001 pengine[15917]: notice: Unfencing > apache-up002.ring0: node discovery > May 18 10:37:03 apache-up001 pengine[15917]: notice: Unfencing > apache-up003.ring0: node discovery > May 18 10:37:03 apache-up001 pengine[15917]: notice: Start > scsia#011(apache-up001.ring0) > May 18 10:37:03 apache-up001 pengine[15917]: notice: Calculated Transition > 11: /var/lib/pacemaker/pengine/pe-input-95.bz2 > May 18 10:37:03 apache-up001 crmd[15918]: notice: Executing on fencing > operation (11) on apache-up003.ring0 (timeout=6) > May 18 10:37:03 apache-up001 crmd[15918]: notice: Initiating action 9: > probe_complete probe_complete-apache-up003.ring0 on apache-up003.ring0 - no > waiting > May 18 10:37:03 apache-up001 crmd[15918]: notice: Executing on fencing > operation (8) on apache-up002.ring0 (timeout=6) > May 18 10:37:03 apache-up001 crmd[15918]: notice: Initiating action 6: > probe_complete probe_complete-apache-up002.ring0 on apache-up002.ring0 - no > waiting > May 18 10:37:03 apache-up001 crmd[15918]: notice: Executing on fencing > operation (5) on apache-up001.ring0 (timeout=6) > May 18 10:37:03 apache-up001 stonith-ng[15914]: notice: Client > crmd.15918.697c495e wants to fence (on) 'apache-up003.ring0' with device > '(any)' > May 18 10:37:03 apache-up001 stonith-ng[15914]: notice: Initiating remote > operation on for apache-up003.ring0: 0599387e-0a30-4e1b-b641-adea5ba2a4ad (0) > May 18 10:37:03 apache-up001 stonith-ng[15914]: notice: Client > crmd.15918.697c495e wants to fence (on) 'apache-up002.ring0' with device > '(any)' > May 18 10:37:03 apache-up001 stonith-ng[15914]: notice: Initiating remote > operation on for apache-up002.ring0: 76aba815-280e-491a-bd17-40776c8169e9 (0) > May 18 10:37:03 apache-up001 crmd[15918]: notice: Initiating action 3: > probe_complete probe_complete-apache-up001.ring0 on apache-up001.ring0 > (local) - no waiting > May 18 10:37:03 apache-up001 stonith-ng[15914]: notice: Client > crmd.15918.697c495e wants to fence (on) 'apache-up001.ring0' with device > '(any)' > May 18 10:37:03 apache-up001 stonith-ng[15914]: notice: Initiating remote > operation on for apache-up001.ring0: e50d7e16-9578-4964-96a3-7b36bdcfba46 (0) > May 18 10:37:03 apache-up001 stonith-ng[15914]: notice: Couldn't find anyone > to fence (on) apache-up003.ring0 with any device > May 18 10:37:03 apache-up001 stonith-ng[15914]: notice: Couldn't find anyone > to fence (on) apache-up002.ring0 with any device > May 18 10:37:03 apache-up001 ston
Re: [ClusterLabs] unable to start fence_scsi
Hi Ken, thank you for the reply I tried as you suggested, and now the stonith devices tries to start but fails. I tried this pcs stonith create scsi fence_scsi pcmk_host_list="apache-up001.ring0 apache-up002.ring0 apache-up003.ring0" pcmk_host_map="apache-up001.ring1=apache-up001.ring0; apache-up002.ring1=apache-up002.ring0; apache-up003.ring1=apache-up003.ring0" pcmk_reboot_action="off" devices="/dev/mapper/36001405973e201b3fdb4a999175b942f" meta provides="unfencing" op monitor interval=60s and even this, adding pcmk_monitor_action="metadata” as suggested in a post on RH knowledge base (even if the error was quite different) pcs stonith create scsi fence_scsi pcmk_host_list="apache-up001.ring0 apache-up002.ring0 apache-up003.ring0" pcmk_host_map="apache-up001.ring1=apache-up001.ring0; apache-up002.ring1=apache-up002.ring0; apache-up003.ring1=apache-up003.ring0" pcmk_reboot_action="off" devices="/dev/mapper/36001405973e201b3fdb4a999175b942f" meta provides="unfencing" pcmk_monitor_action="metadata" op monitor interval=60s I’m using CentOS 7.2, pacemaker-1.1.13-10 resource-agents-3.9.5-54 and fence-agents-scsi-4.0.11-27 the error message are Couldn't find anyone to fence (on) apache-up003.ring0 with any device anderror: Operation on of apache-up003.ring0 by for crmd.15918@apache-up001.ring0.0599387e: No such device Thanks Marco May 18 10:37:03 apache-up001 crmd[15918]: notice: State transition S_IDLE -> S_POLICY_ENGINE [ input=I_PE_CALC cause=C_FSA_INTERNAL origin=abort_transition_graph ] May 18 10:37:03 apache-up001 pengine[15917]: notice: On loss of CCM Quorum: Ignore May 18 10:37:03 apache-up001 pengine[15917]: notice: Unfencing apache-up001.ring0: node discovery May 18 10:37:03 apache-up001 pengine[15917]: notice: Unfencing apache-up002.ring0: node discovery May 18 10:37:03 apache-up001 pengine[15917]: notice: Unfencing apache-up003.ring0: node discovery May 18 10:37:03 apache-up001 pengine[15917]: notice: Start scsia#011(apache-up001.ring0) May 18 10:37:03 apache-up001 pengine[15917]: notice: Calculated Transition 11: /var/lib/pacemaker/pengine/pe-input-95.bz2 May 18 10:37:03 apache-up001 crmd[15918]: notice: Executing on fencing operation (11) on apache-up003.ring0 (timeout=6) May 18 10:37:03 apache-up001 crmd[15918]: notice: Initiating action 9: probe_complete probe_complete-apache-up003.ring0 on apache-up003.ring0 - no waiting May 18 10:37:03 apache-up001 crmd[15918]: notice: Executing on fencing operation (8) on apache-up002.ring0 (timeout=6) May 18 10:37:03 apache-up001 crmd[15918]: notice: Initiating action 6: probe_complete probe_complete-apache-up002.ring0 on apache-up002.ring0 - no waiting May 18 10:37:03 apache-up001 crmd[15918]: notice: Executing on fencing operation (5) on apache-up001.ring0 (timeout=6) May 18 10:37:03 apache-up001 stonith-ng[15914]: notice: Client crmd.15918.697c495e wants to fence (on) 'apache-up003.ring0' with device '(any)' May 18 10:37:03 apache-up001 stonith-ng[15914]: notice: Initiating remote operation on for apache-up003.ring0: 0599387e-0a30-4e1b-b641-adea5ba2a4ad (0) May 18 10:37:03 apache-up001 stonith-ng[15914]: notice: Client crmd.15918.697c495e wants to fence (on) 'apache-up002.ring0' with device '(any)' May 18 10:37:03 apache-up001 stonith-ng[15914]: notice: Initiating remote operation on for apache-up002.ring0: 76aba815-280e-491a-bd17-40776c8169e9 (0) May 18 10:37:03 apache-up001 crmd[15918]: notice: Initiating action 3: probe_complete probe_complete-apache-up001.ring0 on apache-up001.ring0 (local) - no waiting May 18 10:37:03 apache-up001 stonith-ng[15914]: notice: Client crmd.15918.697c495e wants to fence (on) 'apache-up001.ring0' with device '(any)' May 18 10:37:03 apache-up001 stonith-ng[15914]: notice: Initiating remote operation on for apache-up001.ring0: e50d7e16-9578-4964-96a3-7b36bdcfba46 (0) May 18 10:37:03 apache-up001 stonith-ng[15914]: notice: Couldn't find anyone to fence (on) apache-up003.ring0 with any device May 18 10:37:03 apache-up001 stonith-ng[15914]: notice: Couldn't find anyone to fence (on) apache-up002.ring0 with any device May 18 10:37:03 apache-up001 stonith-ng[15914]: error: Operation on of apache-up003.ring0 by for crmd.15918@apache-up001.ring0.0599387e: No such device May 18 10:37:03 apache-up001 stonith-ng[15914]: error: Operation on of apache-up002.ring0 by for crmd.15918@apache-up001.ring0.76aba815: No such device May 18 10:37:03 apache-up001 stonith-ng[15914]: notice: Couldn't find anyone to fence (on) apache-up001.ring0 with any device May 18 10:37:03 apache-up001 crmd[15918]: notice: Stonith operation 5/11:11:0:8248cebf-c198-4ff2-bd43-7415533ce50f: No such device (-19) May 18 10:37:03 apache-up001 stonith-ng[15914]: error: Operation on of apache-up001.ring0 by for crmd.15918@apache-up001.ri
Re: [ClusterLabs] SAN with drbd and pacemaker
"With 20 disk of 4TB you have a total capacity of 80TB. If you run all of them as RAID6 then you have a total of 72TB." And that's the point! I'm trying to understand if I can create more RAID6 arrays and how my controller handles disk failures in that case. First I think we need to clarify terminology related to Megaraid Storage Manager and for this reason I attach here a screenshot -> (phyical drives -> http://pasteboard.co/NC3O60x.pngand logical drives -> http://pasteboard.co/NC8DLcM.png ) So, reading this -> http://www.cisco.com/c/dam/en/us/td/docs/unified_computing/ucs/3rd-party/lsi/mrsas/userguide/LSI_MR_SAS_SW_UG.pdf (page 41) I think I have a RAID (6 in my case) array as a drive group and a volume as a Virtual Drive. If this is right, I should discover how much RAID array my controller supports. Actually I have 20 disks, but I can add more. However, reducing rebuild array time is my goal, so I think that create a virtual drive for each drive group is the right way. Please give me some advises Thanks 2015-09-18 13:02 GMT+02:00 Kai Dupke <kdu...@suse.com>: > On 09/18/2015 09:28 AM, Marco Marino wrote: > > Can you explain me this? 16 volumes? > > > With 20 disk of 4TB you have a total capacity of 80TB. If you run all of > them as RAID6 then you have a total of 72TB. > > If you ask your controller to create a 8TB volume, this volume is spread > across all the 20 disk. As 2 stripes are used for parity, you have > 20-2=18 data stripes per volume. This makes each stripe 444G big, > leaving 3500G free for other volumes. > > If you fill up the remaining 3500G with volumes the same way, you get 8 > additional volumes (OK, the last volume is <8TB then). > > In total you have 9 volumes then, each disk has data/parity on all of > these volumes. > > 9x8=72, voila! > > If a disk error appear and the controller marks the disk dead then all 9 > volumes are affected. > > With 20 6TB/8TB drives, you just get more 8TB volumes using this way. > > What would of course reduce the risk is to always use <20 disk in one > raid6 volume, so not each disk serves all volumes. > > Another issue is about performance, not every RAID controller performs > best with 20 drives. Adaptec recommends an odd number of drives, with 7 > or 9 drives performs best AFAIK. > > So you could make volume 1 on disks 1-9, volume 2 on disk 2-10, volume 3 > on disk 3-11 etc. etc. > > Or consider using some combination of RAID6 and RAID1, but this gives > you way less available disk size (and no, I have no calculation handy on > the chance for failure for RAID6 vs. RAID15 vs. RAID16) > > greetings kai > > > > > > Thank you > > > > > > > > 2015-09-17 15:54 GMT+02:00 Kai Dupke <kdu...@suse.com>: > > > >> On 09/17/2015 09:44 AM, Marco Marino wrote: > >>> Hi, I have 2 servers supermicro lsi 2108 with many disks (80TB) and I'm > >>> trying to build a SAN with drbd and pacemaker. I'm studying, but I have > >> no > >>> experience on large array of disks with drbd and pacemaker, so I have > >> some > >>> questions: > >>> > >>> I'm using MegaRAID Storage Manager to create virtual drives. Each > virtual > >>> drive is a device on linux (eg /dev/sdb, /dev/sdc.), so my first > >>> question is: it's a good idea to create virtual drive of 8 TB (max)? > I'm > >>> thinking to rebuild array time in case of disk failure (about 1 day > for 8 > >> > >> It depends on your disks and RAID level. If one disk fails the content > >> of this disk has to be recreated by either copying (all RAID levels with > >> some RAID 1 included) or calculating (all with no RAID1 included), in > >> the later case all disks get really stressed. > >> > >> If you run 20x4TB disks as RAID6, then an 8TB volume is only ~500G per > >> disk. However, if one disk fails, then all the other 15 volumes this > >> disk handles are broken, too. (BTW, most raid controller can handle > >> multiple stripes per disk, but usually only a handful) In such case the > >> complete 4TB of the broken disk has to be recovered, affecting all 16 > >> volumes. > >> > >> On the other side, if you use 4x5x4TB as 4x 12TB RAID6, a broken disk > >> only affects one of 4 volumes - but at the cost of more disks needed. > >> > >> You can do the similar calculation based on RAID16/15. > >> > >> The only reason I see to create small slices is to make them fit on > >> smaller replacement disks, which might be more easily available/payable > >>
Re: [ClusterLabs] SAN with drbd and pacemaker
ok, first if all, thank you for your answer. This is acomplicated task and I cannot found many guides (if you have are welcome). I'm using RAID6 and I have 20 disks of 4TB each. In RAID6 space efficiency is 1-2/n, so a solution for small Virtual Drive could be 4 or 5 disks. If I use 4 disks I will have (4*4) * (1-2/4) = 8 TB of effective space. Instead, if I use 5 disks, I will have (5*4) * (1-2/5) = 12TB of effective space. Space efficiency is not a primary goal for me, I'm trying to reduce time of rebuilding when a disk fails (and performance improvement!). "If you run 20x4TB disks as RAID6, then an 8TB volume is only ~500G per disk. However, if one disk fails, then all the other 15 volumes this disk handles are broken, too. (BTW, most raid controller can handle multiple stripes per disk, but usually only a handful) In such case the complete 4TB of the broken disk has to be recovered, affecting all 16 volumes." Can you explain me this? 16 volumes? Thank you 2015-09-17 15:54 GMT+02:00 Kai Dupke <kdu...@suse.com>: > On 09/17/2015 09:44 AM, Marco Marino wrote: > > Hi, I have 2 servers supermicro lsi 2108 with many disks (80TB) and I'm > > trying to build a SAN with drbd and pacemaker. I'm studying, but I have > no > > experience on large array of disks with drbd and pacemaker, so I have > some > > questions: > > > > I'm using MegaRAID Storage Manager to create virtual drives. Each virtual > > drive is a device on linux (eg /dev/sdb, /dev/sdc.), so my first > > question is: it's a good idea to create virtual drive of 8 TB (max)? I'm > > thinking to rebuild array time in case of disk failure (about 1 day for 8 > > It depends on your disks and RAID level. If one disk fails the content > of this disk has to be recreated by either copying (all RAID levels with > some RAID 1 included) or calculating (all with no RAID1 included), in > the later case all disks get really stressed. > > If you run 20x4TB disks as RAID6, then an 8TB volume is only ~500G per > disk. However, if one disk fails, then all the other 15 volumes this > disk handles are broken, too. (BTW, most raid controller can handle > multiple stripes per disk, but usually only a handful) In such case the > complete 4TB of the broken disk has to be recovered, affecting all 16 > volumes. > > On the other side, if you use 4x5x4TB as 4x 12TB RAID6, a broken disk > only affects one of 4 volumes - but at the cost of more disks needed. > > You can do the similar calculation based on RAID16/15. > > The only reason I see to create small slices is to make them fit on > smaller replacement disks, which might be more easily available/payable > at time of error (but now we are entering a more low cost area where > usually SAN and DRBD do not take place). > > greetings > Kai Dupke > Senior Product Manager > Server Product Line > -- > Sell not virtue to purchase wealth, nor liberty to purchase power. > Phone: +49-(0)5102-9310828 Mail: kdu...@suse.com > Mobile: +49-(0)173-5876766 WWW: www.suse.com > > SUSE Linux GmbH - Maxfeldstr. 5 - 90409 Nuernberg (Germany) > GF:Felix Imendörffer,Jane Smithard,Graham Norton,HRB 21284 (AG Nürnberg) > > ___ > Users mailing list: Users@clusterlabs.org > http://clusterlabs.org/mailman/listinfo/users > > Project Home: http://www.clusterlabs.org > Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf > Bugs: http://bugs.clusterlabs.org > ___ Users mailing list: Users@clusterlabs.org http://clusterlabs.org/mailman/listinfo/users Project Home: http://www.clusterlabs.org Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf Bugs: http://bugs.clusterlabs.org
