[Linux-HA] ipvssync/ldirectord
Hello list, what is the current status/best practise for ipvsadm with ldirectord? should ipvsadm --both defined as a clone resource? which constraints or order should be taken? currently I only run ldirectord on the active VIP. I want to use LVS with DR. Thanks for your reply in advance. Tom. ___ Linux-HA mailing list Linux-HA@lists.linux-ha.org http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
Re: [Linux-HA] ipvssync/ldirectord
Am Donnerstag, 25. Februar 2010 10:19:08 schrieb bi...@antworte.me: > Hello list, > > what is the current status/best practise for ipvsadm with ldirectord? > > should ipvsadm --both defined as a clone resource? > which constraints or order should be taken? > currently I only run ldirectord on the active VIP. > > I want to use LVS with DR. > > Thanks for your reply in advance. > > Tom. I have ipvsadm replication started by the init system and ldirectord managed by the cluster (pacemaker). Works like a charm. Michael. -- Dr. Michael Schwartzkopff MultiNET Services GmbH Addresse: Bretonischer Ring 7; 85630 Grasbrunn; Germany Tel: +49 - 89 - 45 69 11 0 Fax: +49 - 89 - 45 69 11 21 mob: +49 - 174 - 343 28 75 mail: mi...@multinet.de web: www.multinet.de Sitz der Gesellschaft: 85630 Grasbrunn Registergericht: Amtsgericht München HRB 114375 Geschäftsführer: Günter Jurgeneit, Hubert Martens --- PGP Fingerprint: F919 3919 FF12 ED5A 2801 DEA6 AA77 57A4 EDD8 979B Skype: misch42 ___ Linux-HA mailing list Linux-HA@lists.linux-ha.org http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
Re: [Linux-HA] ipvssync/ldirectord
> I have ipvsadm replication started by the init system and ldirectord > managed > by the cluster (pacemaker). Works like a charm. > > Michael. Sounds good. IMHO the ipvsadm rules are controlled by ldirectord.cf. So my ipvsadm.rules are currently empty. which are the minimum ipvsadm.rules to work in your charming setup? e.g. for 3 or 4 nodes? Thanks for your reply! Tom. ___ Linux-HA mailing list Linux-HA@lists.linux-ha.org http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
Re: [Linux-HA] ipvssync/ldirectord
Am Donnerstag, 25. Februar 2010 10:34:17 schrieb bi...@antworte.me: > > I have ipvsadm replication started by the init system and ldirectord > > managed > > by the cluster (pacemaker). Works like a charm. > > > > Michael. > > Sounds good. IMHO the ipvsadm rules are controlled by ldirectord.cf. > So my ipvsadm.rules are currently empty. > which are the minimum ipvsadm.rules to work in your charming setup? > e.g. for 3 or 4 nodes? > > Thanks for your reply! > > Tom. Don't know. Rules are controlled by ldirectord. So I don't deal with the rules direct. Minimum needed: One virtual server and at least one real server. Maximum: Lots! -- Dr. Michael Schwartzkopff MultiNET Services GmbH Addresse: Bretonischer Ring 7; 85630 Grasbrunn; Germany Tel: +49 - 89 - 45 69 11 0 Fax: +49 - 89 - 45 69 11 21 mob: +49 - 174 - 343 28 75 mail: mi...@multinet.de web: www.multinet.de Sitz der Gesellschaft: 85630 Grasbrunn Registergericht: Amtsgericht München HRB 114375 Geschäftsführer: Günter Jurgeneit, Hubert Martens --- PGP Fingerprint: F919 3919 FF12 ED5A 2801 DEA6 AA77 57A4 EDD8 979B Skype: misch42 ___ Linux-HA mailing list Linux-HA@lists.linux-ha.org http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
Re: [Linux-HA] ipvssync/ldirectord
>> Sounds good. IMHO the ipvsadm rules are controlled by ldirectord.cf. >> So my ipvsadm.rules are currently empty. >> which are the minimum ipvsadm.rules to work in your charming setup? >> e.g. for 3 or 4 nodes? >> >> Thanks for your reply! >> >> Tom. > > Don't know. Rules are controlled by ldirectord. So I don't deal with the > rules > direct. > > Minimum needed: One virtual server and at least one real server. > Maximum: Lots! Does this mean - it doesnt matter what is in the ipvsadm.rules? It just needs to have the ipvsadm started ? ___ Linux-HA mailing list Linux-HA@lists.linux-ha.org http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
Re: [Linux-HA] ipvssync/ldirectord
Am Donnerstag, 25. Februar 2010 11:13:41 schrieb bi...@antworte.me: > >> Sounds good. IMHO the ipvsadm rules are controlled by ldirectord.cf. > >> So my ipvsadm.rules are currently empty. > >> which are the minimum ipvsadm.rules to work in your charming setup? > >> e.g. for 3 or 4 nodes? > >> > >> Thanks for your reply! > >> > >> Tom. > > > > Don't know. Rules are controlled by ldirectord. So I don't deal with the > > rules > > direct. > > > > Minimum needed: One virtual server and at least one real server. > > Maximum: Lots! > > Does this mean - it doesnt matter what is in the ipvsadm.rules? > It just needs to have the ipvsadm started ? ipvsadm is responsible for the state sync. ldirectord for the rules. -- Dr. Michael Schwartzkopff MultiNET Services GmbH Addresse: Bretonischer Ring 7; 85630 Grasbrunn; Germany Tel: +49 - 89 - 45 69 11 0 Fax: +49 - 89 - 45 69 11 21 mob: +49 - 174 - 343 28 75 mail: mi...@multinet.de web: www.multinet.de Sitz der Gesellschaft: 85630 Grasbrunn Registergericht: Amtsgericht München HRB 114375 Geschäftsführer: Günter Jurgeneit, Hubert Martens --- PGP Fingerprint: F919 3919 FF12 ED5A 2801 DEA6 AA77 57A4 EDD8 979B Skype: misch42 ___ Linux-HA mailing list Linux-HA@lists.linux-ha.org http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
Re: [Linux-HA] ipvssync/ldirectord
Am 25.02.10 11:13, schrieb bi...@antworte.me: > >>> Sounds good. IMHO the ipvsadm rules are controlled by ldirectord.cf. >>> So my ipvsadm.rules are currently empty. >>> which are the minimum ipvsadm.rules to work in your charming setup? >>> e.g. for 3 or 4 nodes? >>> >>> Thanks for your reply! >>> >>> Tom. >>> >> Don't know. Rules are controlled by ldirectord. So I don't deal with the >> rules >> direct. >> >> Minimum needed: One virtual server and at least one real server. >> Maximum: Lots! >> > Does this mean - it doesnt matter what is in the ipvsadm.rules? > It just needs to have the ipvsadm started ? Yes, this is exactly what it means, ipvsadm rules are controlled via ldirectord in most cases. Most people which to have their ressources actively monitored and reacted on failed hosts / services . So ldirectord enables / disables hosts / services according to the configured test rule(s). ipvsadm.rules was meant to have static rules reloaded on reboot - without something else controlling them. kind regards, Malte Geierhos ___ Linux-HA mailing list Linux-HA@lists.linux-ha.org http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
Re: [Linux-HA] ipvssync/ldirectord
-Original Message- From: linux-ha-boun...@lists.linux-ha.org [mailto:linux-ha-boun...@lists.linux-ha.org] On Behalf Of Malte Geierhos Sent: 25 February 2010 10:18 To: linux-ha@lists.linux-ha.org Subject: Re: [Linux-HA] ipvssync/ldirectord Am 25.02.10 11:13, schrieb bi...@antworte.me: > >>> Sounds good. IMHO the ipvsadm rules are controlled by ldirectord.cf. >>> So my ipvsadm.rules are currently empty. >>> which are the minimum ipvsadm.rules to work in your charming setup? >>> e.g. for 3 or 4 nodes? >>> >>> Thanks for your reply! >>> >>> Tom. >>> >> Don't know. Rules are controlled by ldirectord. So I don't deal with the >> rules >> direct. >> >> Minimum needed: One virtual server and at least one real server. >> Maximum: Lots! >> > Does this mean - it doesnt matter what is in the ipvsadm.rules? > It just needs to have the ipvsadm started ? Yes, this is exactly what it means, ipvsadm rules are controlled via ldirectord in most cases. Most people which to have their ressources actively monitored and reacted on failed hosts / services . So ldirectord enables / disables hosts / services according to the configured test rule(s). ipvsadm.rules was meant to have static rules reloaded on reboot - without something else controlling them. kind regards, Malte Geierhos -- Previously I didn't have ipvsadm chkconfig'd on, ldirectord just loaded rules itself. Now I have the ipvsadm init script start on boot, just so it loads the ipvsadm.rules so I can run the sync daemon. - Darren ___ Linux-HA mailing list Linux-HA@lists.linux-ha.org http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
Re: [Linux-HA] ipvssync/ldirectord
>> Does this mean - it doesnt matter what is in the ipvsadm.rules? >> It just needs to have the ipvsadm started ? > > ipvsadm is responsible for the state sync. > ldirectord for the rules. So -- if i have a init script, which does a ipvsadm --start-daemon=master --syncid=254 --mcast-interface=eth0 ipvsadm --start-daemon=backup --syncid=254 --mcast-interface=eth0 on all nodes, I have everything to get the state-table synced to the other nodes, and the rules are controlled by ldirectord? Right? Thanks for your reply in advance. Tom. ___ Linux-HA mailing list Linux-HA@lists.linux-ha.org http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
[Linux-HA] Filesystem starting at the wrong time
Hi there, I have a config setup in Ubuntu 9.10 with Pacemaker 1.0.5 here it is node $id="76df8738-d836-4a95-afe3-732bc5cc7b66" ewadata2 \ attributes standby="off" node $id="fd56770d-eef2-47e5-97aa-813ae1e71d7f" ewadata1 \ attributes standby="off" primitive IP ocf:heartbeat:IPaddr \ params ip="10.65.68.245" \ meta is-managed="true" target-role="Started" primitive drbd_r0 ocf:heartbeat:drbd \ params drbd_resource="r0" \ op monitor interval="9s" role="Master" timeout="20s" \ op monitor interval="10s" role="Slave" timeout="20s" primitive fs-drbd_r0 ocf:heartbeat:Filesystem \ params fstype="ext4" directory="/data" device="/dev/drbd0" \ op monitor interval="5s" ms ms-drbd_r0 drbd_r0 \ meta clone-max="2" notify="true" globally-unique="false" location fs-drbd fs-drbd_r0 \ rule $id="fs-drbd-rule" inf: #uname eq ewadata1 location l_prefer_IP IP \ rule $id="l_prefer_IP-rule" inf: #uname eq ewadata1 location ms-drbd-master ms-drbd_r0 \ rule $id="ms-drbd-master-rule" $role="master" inf: #uname eq ewadata1 order IP_before_ms-drbd inf: IP ms-drbd_r0:start order ms-drbd_before_fs-drbd inf: ms-drbd_r0:promote fs-drbd:start property $id="cib-bootstrap-options" \ dc-version="1.0.5-3840e6b5a305ccb803d29b468556739e75532d56" \ cluster-infrastructure="Heartbeat" \ no-quorum-policy="ignore" \ last-lrm-refresh="1267094619" \ stonith-enabled="false" I've set the order that IP must start before ms-drbd does it's promoting and that ms-drbd has to be promtoted before it start the filesystem. But however if I do a switch with crm node standby ewadata1 the filesystem tries to start itself imediately not waiting for the IP to come up on ewadata2. Did I miss something in my config? Thx for any hints. ___ Linux-HA mailing list Linux-HA@lists.linux-ha.org http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
Re: [Linux-HA] ipvssync/ldirectord
Am Donnerstag, 25. Februar 2010 11:37:24 schrieb bi...@antworte.me: > >> Does this mean - it doesnt matter what is in the ipvsadm.rules? > >> It just needs to have the ipvsadm started ? > > > > ipvsadm is responsible for the state sync. > > ldirectord for the rules. > > So -- if i have a init script, which does a > > ipvsadm --start-daemon=master --syncid=254 --mcast-interface=eth0 > ipvsadm --start-daemon=backup --syncid=254 --mcast-interface=eth0 > > on all nodes, I have everything to get the state-table synced to > the other nodes, and the rules are controlled by ldirectord? > Right? > > Thanks for your reply in advance. > > Tom. Yes. -- Dr. Michael Schwartzkopff MultiNET Services GmbH Addresse: Bretonischer Ring 7; 85630 Grasbrunn; Germany Tel: +49 - 89 - 45 69 11 0 Fax: +49 - 89 - 45 69 11 21 mob: +49 - 174 - 343 28 75 mail: mi...@multinet.de web: www.multinet.de Sitz der Gesellschaft: 85630 Grasbrunn Registergericht: Amtsgericht München HRB 114375 Geschäftsführer: Günter Jurgeneit, Hubert Martens --- PGP Fingerprint: F919 3919 FF12 ED5A 2801 DEA6 AA77 57A4 EDD8 979B Skype: misch42 ___ Linux-HA mailing list Linux-HA@lists.linux-ha.org http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
Re: [Linux-HA] Filesystem starting at the wrong time
Am Donnerstag, 25. Februar 2010 11:54:57 schrieb Frank Lazzarini: > Hi there, I have a config setup in Ubuntu 9.10 with Pacemaker 1.0.5 here > it is > > node $id="76df8738-d836-4a95-afe3-732bc5cc7b66" ewadata2 \ > attributes standby="off" > node $id="fd56770d-eef2-47e5-97aa-813ae1e71d7f" ewadata1 \ > attributes standby="off" > > primitive IP ocf:heartbeat:IPaddr \ > params ip="10.65.68.245" \ > meta is-managed="true" target-role="Started" > primitive drbd_r0 ocf:heartbeat:drbd \ > params drbd_resource="r0" \ > op monitor interval="9s" role="Master" timeout="20s" \ > op monitor interval="10s" role="Slave" timeout="20s" > primitive fs-drbd_r0 ocf:heartbeat:Filesystem \ > params fstype="ext4" directory="/data" device="/dev/drbd0" \ > op monitor interval="5s" > > ms ms-drbd_r0 drbd_r0 \ > meta clone-max="2" notify="true" globally-unique="false" > > location fs-drbd fs-drbd_r0 \ > rule $id="fs-drbd-rule" inf: #uname eq ewadata1 > location l_prefer_IP IP \ > rule $id="l_prefer_IP-rule" inf: #uname eq ewadata1 > location ms-drbd-master ms-drbd_r0 \ > rule $id="ms-drbd-master-rule" $role="master" inf: #uname eq ewadata1 Do you have any special reason why do you want to have the service run on that machine? Are both machines not identically? I'd throw these rules out. > order IP_before_ms-drbd inf: IP ms-drbd_r0:start > order ms-drbd_before_fs-drbd inf: ms-drbd_r0:promote fs-drbd:start order ms-drbd_before_fs inf: ms-drbd_r0:promote fs-drbd:start Also make a colocation colocation fs_with_drbd inf: fs-drbd: Started ms-drbd-r0:Master Make a group out of the fs-drbd and the IP. Greetings, -- Dr. Michael Schwartzkopff MultiNET Services GmbH Addresse: Bretonischer Ring 7; 85630 Grasbrunn; Germany Tel: +49 - 89 - 45 69 11 0 Fax: +49 - 89 - 45 69 11 21 mob: +49 - 174 - 343 28 75 mail: mi...@multinet.de web: www.multinet.de Sitz der Gesellschaft: 85630 Grasbrunn Registergericht: Amtsgericht München HRB 114375 Geschäftsführer: Günter Jurgeneit, Hubert Martens --- PGP Fingerprint: F919 3919 FF12 ED5A 2801 DEA6 AA77 57A4 EDD8 979B Skype: misch42 ___ Linux-HA mailing list Linux-HA@lists.linux-ha.org http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
[Linux-HA] crm configure edit not allowed in batch mode
Hi, i want to edit a resource listed by crm configure show via batch/cmdline. current: group loadbalancer ldirectord vip_1 vip_2 should be changed to: group loadbalancer ldirectord vip_1 vip_2 vip_3 I tried echo "group loadbalancer ldirectord vip_1 vip_2 vip_3" | crm configure edit gp_LVS But there is following message: INFO: no curses support: you won't see colors INFO: edit not allowed in batch mode Btw. where can the EDITOR or PAGER been controlled? Currently it is vi and less Thanks for your reply in advance. Tom. ___ Linux-HA mailing list Linux-HA@lists.linux-ha.org http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
Re: [Linux-HA] crm configure edit not allowed in batch mode
Hi, On Thu, Feb 25, 2010 at 12:45:23PM +0100, bi...@antworte.me wrote: > > Hi, i want to edit a resource listed by crm configure show via > batch/cmdline. > > current: > > group loadbalancer ldirectord vip_1 vip_2 > > should be changed to: > > group loadbalancer ldirectord vip_1 vip_2 vip_3 > > > I tried > > echo "group loadbalancer ldirectord vip_1 vip_2 vip_3" | crm configure > edit gp_LVS > > But there is following message: > > INFO: no curses support: you won't see colors > INFO: edit not allowed in batch mode Edit means invoking editor, but you can't do that if there's no terminal. Right now, there's no stream editor, though that may change. This should do: cat < Btw. > > where can the EDITOR or PAGER been controlled? > Currently it is vi and less # crm options editor the_other_editor # crm options pager the_other_pager You should also save options, if you set them interactively. Thanks, Dejan > Thanks for your reply in advance. > > Tom. > ___ > Linux-HA mailing list > Linux-HA@lists.linux-ha.org > http://lists.linux-ha.org/mailman/listinfo/linux-ha > See also: http://linux-ha.org/ReportingProblems ___ Linux-HA mailing list Linux-HA@lists.linux-ha.org http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
Re: [Linux-HA] crm configure edit not allowed in batch mode
On 2010-02-25 14:01, Dejan Muhamedagic wrote: > This should do: > > cat < delete loadbalancer > group loadbalancer ldirectord vip_1 vip_2 vip_3 > EOF You will also need to insert "commit" before EOF. And of course, no need for cat, "crm configure <". Cheers, Florian signature.asc Description: OpenPGP digital signature ___ Linux-HA mailing list Linux-HA@lists.linux-ha.org http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
Re: [Linux-HA] Command line option to fail back a master/slave resource
Hi, On Wed, Feb 24, 2010 at 10:45:16AM -0800, Bob Schatz wrote: > Dejan, > > > Sorry for the delay in responding. > > Comments inline with [BS] and questions with "QUESTION #" > > > Thanks! > > Bob > > > > - Original Message > From: Dejan Muhamedagic > To: General Linux-HA mailing list > Sent: Fri, February 19, 2010 4:29:01 AM > Subject: Re: [Linux-HA] Command line option to fail back a master/slave > resource > > Hi, > > On Thu, Feb 18, 2010 at 05:18:21PM -0800, Bob Schatz wrote: > > Thanks Marian and Dejan! > > > > I did these steps for fail back: > > > > # crm resource meta SS0 delete target-role > > # crm resource demote ms-SS0 > > # crm resource promote ms-SS0 > > > > I noticed that if I type to fast between the "demote" and the "promote" > > then the ms-SS0 > > ends going MASTER->SLAVE->MASTER on the same node. > > Are you sure that the transition actually takes place? Maybe the > second command aborts it half way. > > BS] Thanks. I think that they are being aborted. I see this in ha-debug > from the DC (log is attached): > > Feb 19 16:43:17 fc12-64-2 crmd: [20062]: debug: te_update_diff: Processing > diff (cib_delete): 0.395.287 -> 0.396.1 (S_IDLE) > Feb 19 16:43:17 fc12-64-2 crmd: [20062]: debug: get_xpath_object: No match > for //cib_update_result//diff-added//crm_config in > /notify/cib_update_result/diff > Feb 19 16:43:17 fc12-64-2 crmd: [20062]: info: abort_transition_graph: > need_abort:59 - Triggered transition abort (complete=1) : Non-status change > Feb 19 16:43:17 fc12-64-2 crmd: [20062]: info: need_abort: Aborting on change > to admin_epoch > Feb 19 16:43:17 fc12-64-2 crmd: [20062]: debug: s_crmd_fsa: Processing > I_PE_CALC: [ state=S_IDLE cause=C_FSA_INTERNAL origin=abort_transition_graph ] > Feb 19 16:43:17 fc12-64-2 crmd: [20062]: info: do_state_transition: State > transition S_IDLE -> S_POLICY_ENGINE [ input=I_PE_CALC cause=C_FSA_INTERNAL > origin=abort_transition_graph ] > Feb 19 16:43:17 fc12-64-2 crmd: [20062]: info: do_state_transition: All 2 > cluster nodes are eligible to run resources. > Feb 19 16:43:17 fc12-64-2 crmd: [20062]: debug: do_fsa_action: actions:trace: > // A_DC_TIMER_STOP > Feb 19 16:43:17 fc12-64-2 crmd: [20062]: debug: do_fsa_action: actions:trace: > // A_INTEGRATE_TIMER_STOP > Feb 19 16:43:17 fc12-64-2 crmd: [20062]: debug: do_fsa_action: actions:trace: > // A_FINALIZE_TIMER_STOP > Feb 19 16:43:17 fc12-64-2 crmd: [20062]: debug: do_fsa_action: actions:trace: > // A_PE_INVOKE > Feb 19 16:43:17 fc12-64-2 crmd: [20062]: info: do_pe_invoke: Query 188: > Requesting the current CIB: S_POLICY_ENGINE > Feb 19 16:43:17 fc12-64-2 cib: [20058]: info: log_data_element: cib:diff: - > > Feb 19 16:43:17 fc12-64-2 cib: [20058]: info: log_data_element: cib:diff: - > > Feb 19 16:43:17 fc12-64-2 cib: [20058]: info: log_data_element: cib:diff: - > > Feb 19 16:43:17 fc12-64-2 cib: [20058]: info: log_data_element: cib:diff: - > > Feb 19 16:43:17 fc12-64-2 cib: [20058]: info: log_data_element: cib:diff: - > > Feb 19 16:43:17 fc12-64-2 cib: [20058]: info: log_data_element: cib:diff: - > > Feb 19 16:43:17 fc12-64-2 cib: [20058]: info: log_data_element: cib:diff: - > value="Started" __crm_diff_marker__="removed:top" /> > Feb 19 16:43:17 fc12-64-2 cib: [20058]: info: log_data_element: cib:diff: - > > Feb 19 16:43:17 fc12-64-2 cib: [20058]: info: log_data_element: cib:diff: - > > Feb 19 16:43:17 fc12-64-2 cib: [20058]: info: log_data_element: cib:diff: - > > Feb 19 16:43:17 fc12-64-2 cib: [20058]: info: log_data_element: cib:diff: - > > Feb 19 16:43:17 fc12-64-2 cib: [20058]: info: log_data_element: cib:diff: - > > Feb 19 16:43:17 fc12-64-2 cib: [20058]: info: log_data_element: cib:diff: - > > Feb 19 16:43:17 fc12-64-2 cib: [20058]: info: log_data_element: cib:diff: + > > > and this: > > Feb 19 16:43:21 fc12-64-2 lrmd: [20059]: info: rsc:SS0:1:112: demote > Feb 19 16:43:21 fc12-64-2 crmd: [20062]: debug: run_graph: Transition 13 > (Complete=6, Pending=1, Fired=1, Skipped=0, Incomplete=6, > Source=/var/lib/pengine/pe-input-0.bz2): In-progress > Feb 19 16:43:21 fc12-64-2 crmd: [20062]: debug: te_update_diff: Processing > diff (cib_modify): 0.397.4 -> 0.398.1 (S_TRANSITION_ENGINE) > Feb 19 16:43:21 fc12-64-2 crmd: [20062]: debug: get_xpath_object: No match > for //cib_update_result//diff-added//crm_config in > /notify/cib_update_result/diff > Feb 19 16:43:21 fc12-64-2 crmd: [20062]: info: abort_transition_graph: > need_abort:59 - Triggered transition abort (complete=0) : Non-status change > Feb 19 16:43:21 fc12-64-2 crmd: [20062]: info: update_abort_priority: Abort > priority upgraded from 0 to 100 > Feb 19 16:43:21 fc12-64-2 crmd: [20062]: info: update_abort_priority: Abort > action done superceeded by restart > Feb 19 16:43:21 fc12-64-2 crmd: [20062]: info: need_abort: Aborting on change > to admin_epoch > Feb 19 16:43:21 fc12-64-2 crmd: [20
Re: [Linux-HA] crm configure edit not allowed in batch mode
On Thu, Feb 25, 2010 at 04:17:32PM +0100, Florian Haas wrote: > On 2010-02-25 14:01, Dejan Muhamedagic wrote: > > This should do: > > > > cat < > delete loadbalancer > > group loadbalancer ldirectord vip_1 vip_2 vip_3 > > EOF > > You will also need to insert "commit" before EOF. And of course, no need > for cat, "crm configure < into a text file, and do "crm configure load update ". Thanks Florian for setting me straight. IIRC, "load update -" reads from stdin so it could also be: echo group loadbalancer ldirectord vip_1 vip_2 vip_3 | crm configure load update - Cheers, Dejan > Cheers, > Florian > > ___ > Linux-HA mailing list > Linux-HA@lists.linux-ha.org > http://lists.linux-ha.org/mailman/listinfo/linux-ha > See also: http://linux-ha.org/ReportingProblems ___ Linux-HA mailing list Linux-HA@lists.linux-ha.org http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
Re: [Linux-HA] Filesystem starting at the wrong time
Thanks a bunch Michael, I want to share my current config ... which is working just perfectly like I wanted it to .. with everyonemaybe it helps someone. :) (note my uuids changed since my last post, and cli-prefer-clustergroup was created automatically when I created the group) node $id="3e440e81-fdeb-4cb5-b41d-f91b982ead58" ewadata1 \ attributes standby="off" node $id="c5a24794-65a1-4b8a-9b1c-7d56a43accce" ewadata2 primitive IP ocf:heartbeat:IPaddr \ params ip="10.65.68.245" primitive drbd0 ocf:heartbeat:drbd \ params drbd_resource="r0" \ op monitor interval="9s" role="Master" timeout="20s" \ op monitor interval="10s" role="Slave" timeout="20s" primitive fs-drbd0 ocf:heartbeat:Filesystem \ params fstype="ext4" directory="/data" device="/dev/drbd0" group clustergroup fs-drbd0 IP ms ms-drbd0 drbd0 \ meta clone-max="2" notify="true" globally-unique="false" location cli-prefer-clustergroup clustergroup \ rule $id="cli-prefer-rule-clustergroup" inf: #uname eq ewadata1 colocation fs_with_drbd inf: clustergroup:Started ms-drbd0:Master order ms-drbd0_before_fs-drbd0 inf: ms-drbd0:promote clustergroup:start property $id="cib-bootstrap-options" \ dc-version="1.0.5-3840e6b5a305ccb803d29b468556739e75532d56" \ cluster-infrastructure="Heartbeat" \ stonith-enabled="false" \ no-quorum-policy="ignore" \ last-lrm-refresh="1267129800" When I put ewadata1 to standby mode everything jumps to ewadata2 and when I turn ewadata1 back on everything stays on ewadata2when I migrate my group resource back to ewadata1 everything switches just fine. One last question is there a way to make that switch back to ewadata1 automatic? (even though it's not really useful). Michael is there a new relase of your O'reilly Book coming? On Thu, Feb 25, 2010 at 12:01 PM, Michael Schwartzkopff wrote: > Am Donnerstag, 25. Februar 2010 11:54:57 schrieb Frank Lazzarini: > > Hi there, I have a config setup in Ubuntu 9.10 with Pacemaker 1.0.5 here > > it is > > > > node $id="76df8738-d836-4a95-afe3-732bc5cc7b66" ewadata2 \ > > attributes standby="off" > > node $id="fd56770d-eef2-47e5-97aa-813ae1e71d7f" ewadata1 \ > > attributes standby="off" > > > > primitive IP ocf:heartbeat:IPaddr \ > > params ip="10.65.68.245" \ > > meta is-managed="true" target-role="Started" > > primitive drbd_r0 ocf:heartbeat:drbd \ > > params drbd_resource="r0" \ > > op monitor interval="9s" role="Master" timeout="20s" \ > > op monitor interval="10s" role="Slave" timeout="20s" > > primitive fs-drbd_r0 ocf:heartbeat:Filesystem \ > > params fstype="ext4" directory="/data" device="/dev/drbd0" \ > > op monitor interval="5s" > > > > ms ms-drbd_r0 drbd_r0 \ > > meta clone-max="2" notify="true" globally-unique="false" > > > > location fs-drbd fs-drbd_r0 \ > > rule $id="fs-drbd-rule" inf: #uname eq ewadata1 > > location l_prefer_IP IP \ > > rule $id="l_prefer_IP-rule" inf: #uname eq ewadata1 > > location ms-drbd-master ms-drbd_r0 \ > > rule $id="ms-drbd-master-rule" $role="master" inf: #uname eq > ewadata1 > > Do you have any special reason why do you want to have the service run on > that > machine? Are both machines not identically? I'd throw these rules out. > > > order IP_before_ms-drbd inf: IP ms-drbd_r0:start > > order ms-drbd_before_fs-drbd inf: ms-drbd_r0:promote fs-drbd:start > > order ms-drbd_before_fs inf: ms-drbd_r0:promote fs-drbd:start > > Also make a colocation > colocation fs_with_drbd inf: fs-drbd: Started ms-drbd-r0:Master > > Make a group out of the fs-drbd and the IP. > > Greetings, > -- > Dr. Michael Schwartzkopff > MultiNET Services GmbH > Addresse: Bretonischer Ring 7; 85630 Grasbrunn; Germany > Tel: +49 - 89 - 45 69 11 0 > Fax: +49 - 89 - 45 69 11 21 > mob: +49 - 174 - 343 28 75 > > mail: mi...@multinet.de > web: www.multinet.de > > Sitz der Gesellschaft: 85630 Grasbrunn > Registergericht: Amtsgericht München HRB 114375 > Geschäftsführer: Günter Jurgeneit, Hubert Martens > > --- > > PGP Fingerprint: F919 3919 FF12 ED5A 2801 DEA6 AA77 57A4 EDD8 979B > Skype: misch42 > ___ > Linux-HA mailing list > Linux-HA@lists.linux-ha.org > http://lists.linux-ha.org/mailman/listinfo/linux-ha > See also: http://linux-ha.org/ReportingProblems > ___ Linux-HA mailing list Linux-HA@lists.linux-ha.org http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
Re: [Linux-HA] IP Power 9258 stonith plugin
Hi Dejan,
Am Mittwoch, 24. Februar 2010 schrieb Dejan Muhamedagic:
> Hi,
>
> On Tue, Feb 23, 2010 at 09:13:19PM +0100, Helmut Weymann wrote:
> > Hi Dejan,
> >
> > many thanks for your comments.
> >
> > I have no idea about the programming standards the linux-ha project uses.
> > I am just a simple user who wants to build a cluster.
> >
> > So, if you want me to apply necessary changes, Please help me. Which name
> > shall I define where?
>
> Not sure what do you mean here. What I meant is that the
> attached file name does not correspond to the plugin name.
What is the right filename?
>
> > On Tue, 23 Feb 2010 09:51:09 +0100, Dejan wrote
> >
> > > Hi,
> > >
> > > On Mon, Feb 22, 2010 at 11:48:17PM +0100, Helmut Weymann wrote:
> > > > Team,
> > > >
> > > > attached is a tested stontih plugin for IP Power 9258 and compatible
> > > > devices. You may want to add it to the product.
> > >
> > > The attached file is misnamed, which doesn't really vouch for
> > > attention to detail. Anyway, here some comments:
> > >
> > > - the way host names are passed is simply dreadful; if there is
> > > no other way then I'd find that a serious obstacle
> >
> > There is a standard html-interface which includes the port names. But, it
> > requires login via post-method. In response to that login the
> > web-interface places a session-cookie and always refers to that cookie
> > whenever a new command is sent.
> >
> > If someone helps me to build the necessary script, I'll do anything
> > necessary to test and finalize that plugin.
> >
> > As a starting point this is the login page:
> >
> > eva:~ # wget -q http://powerswitch/ -O -
> >
> >
> > IP9258
> >
> >
> >
> >
> > function calcResponse()
> > {
> > str = document.login.Username.value +
> > document.login.Password.value +
> > document.login.Challenge.value;
> > document.login.Response.value = hex_md5(str);
> > document.login.Password.value = "";
> > document.login.Challenge.value = "";
> > document.login.submit();
> > }
> >
> >
> >
> >
> >
> >
> >
> >
> > > HEIGHT="93">
> > > COLOR="#FF" SIZE="+1">IP9258 Login
> >
> >
> > Please enter user name and
> > password.
> >
> > User Name:
> > > MAXLENGTH="32">
> >
> > Password:
> > > MAXLENGTH="32">
> >
> > > NAME="Submitbtn" VALUE=" OK "> > NAME="Cancel" VALUE=" Cancel "
> > ONCLICK="document.login.Username.value='';document.login.Password.value='
> >'"> > VALUE="W4nW/...">
> >
> >
> >
> >
> > document.login.Username.focus();
> >
> >
> > login.ScreenWidth.value = screen.width;
> >
> >
> >
>
> Perhaps to save the cookie in /var/run/heartbeat/rsctmp/. wget
> has some cookies options which should help.
The cookie is no problem. Yes, wget and curl can handle the cookie.
The problem is
Re: [Linux-HA] IP Power 9258 stonith plugin
Am Freitag, 26. Februar 2010 schrieb Helmut Weymann:
> Hi Dejan,
>
> Am Mittwoch, 24. Februar 2010 schrieb Dejan Muhamedagic:
> > Hi,
> >
> > On Tue, Feb 23, 2010 at 09:13:19PM +0100, Helmut Weymann wrote:
> > > Hi Dejan,
> > >
> > > many thanks for your comments.
> > >
> > > I have no idea about the programming standards the linux-ha project
> > > uses. I am just a simple user who wants to build a cluster.
> > >
> > > So, if you want me to apply necessary changes, Please help me. Which
> > > name shall I define where?
> >
> > Not sure what do you mean here. What I meant is that the
> > attached file name does not correspond to the plugin name.
>
> What is the right filename?
>
> > > On Tue, 23 Feb 2010 09:51:09 +0100, Dejan wrote
> > >
> > > > Hi,
> > > >
> > > > On Mon, Feb 22, 2010 at 11:48:17PM +0100, Helmut Weymann wrote:
> > > > > Team,
> > > > >
> > > > > attached is a tested stontih plugin for IP Power 9258 and
> > > > > compatible devices. You may want to add it to the product.
> > > >
> > > > The attached file is misnamed, which doesn't really vouch for
> > > > attention to detail. Anyway, here some comments:
> > > >
> > > > - the way host names are passed is simply dreadful; if there is
> > > > no other way then I'd find that a serious obstacle
> > >
> > > There is a standard html-interface which includes the port names. But,
> > > it requires login via post-method. In response to that login the
> > > web-interface places a session-cookie and always refers to that cookie
> > > whenever a new command is sent.
> > >
> > > If someone helps me to build the necessary script, I'll do anything
> > > necessary to test and finalize that plugin.
> > >
> > > As a starting point this is the login page:
> > >
> > > eva:~ # wget -q http://powerswitch/ -O -
> > >
> > >
> > > IP9258
> > >
> > >
> > >
> > >
> > > function calcResponse()
> > > {
> > > str = document.login.Username.value +
> > > document.login.Password.value +
> > > document.login.Challenge.value;
> > > document.login.Response.value = hex_md5(str);
> > > document.login.Password.value = "";
> > > document.login.Challenge.value = "";
> > > document.login.submit();
> > > }
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > > > > HEIGHT="93">
> > > > > COLOR="#FF" SIZE="+1">IP9258 Login
> > >
> > >
> > > Please enter user name and
> > > password.
> > >
> > > User Name:
> > > > > MAXLENGTH="32">
> > >
> > > Password:
> > > > > MAXLENGTH="32">
> > >
> > > > > NAME="Submitbtn" VALUE=" OK "> > > NAME="Cancel" VALUE=" Cancel "
> > > ONCLICK="document.login.Username.value='';document.login.Password.value
> > >=' '"> > > VALUE="W4nW/...">
> > >
> > >
> > >
> > >
> > > document.login.Username.focus();
> > >
> > >
> > > login.ScreenWidth.value = screen.width;
> > >
> > >
> > >
> >
> > Perhaps to save the cookie in /var/run/heartbeat/rsctmp/. wget
> > has some cookies options which should help.
>
> The cookie is no problem. Yes, wget and curl can handle the cookie.
> The problem is
