f



Sun Cluster 3.3 : second node crash when first node crash

Hello,

First : thanks for reading ... :o)

I've just installed a Sun Cluster 3.3 on 2 sol 10 x86 nodes ( 2 VMs on ESXi 5.1 ).

Node Name       Present       Possible       Status
---------       -------       --------       ------
master          1             1              Online
slave           0             0              Online

All seems to be ok for the moment.

I'm doing a test : Hard Power Off the first node, just to check if the quorum switch to the slave.

But few seconds after this : the slave freeze and reset and stays at boot waiting a for a second node. If i don't boot the master, the slave does not want to boot.

It's is normal ? it's very strange and every time i'm doing this test. Why my second node reset when i power off the master !

Thanks you for help. And sorry for my english. I'm french.

Regards
Manu
0
nslmanu (4)
11/11/2012 9:03:27 AM
comp.unix.solaris 26025 articles. 2 followers. Post Follow

6 Replies
604 Views

Similar Articles

[PageSpeed] 57

nslmanu <nslmanu@gmail.com> wrote:
> But few seconds after this : the slave freeze and reset and stays at boot waiting a for a second node. If i don't boot the master, the slave does not want to boot.
> 
> It's is normal ? it's very strange and every time i'm doing this test. Why my second node reset when i power off the master !

What kind of quorum device your cluster have? Cluster needs n+1 quorum in
order to be operational.

what does scstat -q say? or clquorum list -v

Sami
0
11/11/2012 10:55:49 AM
> What kind of quorum device your cluster have? Cluster needs n+1 quorum in
> 
> order to be operational.
> 
> 
> 
> what does scstat -q say? or clquorum list -v
> 
> 
> 
> Sami

Thnaks Sami for your answer.

In fact the goal is for the moment just to have 2 nodes with a apache webserver running with static pages.

So i don't need another device as a disk drive for example. But i don't know if it's really mandatory to have one.

Maybe i just have to add one vote to the second node ?

Tks for your answer.

Regards
Manu
0
nslmanu (4)
11/11/2012 12:53:02 PM
nslmanu <nslmanu@gmail.com> wrote:
> In fact the goal is for the moment just to have 2 nodes with a apache webserver running with static pages.
> 
> So i don't need another device as a disk drive for example. But i don't know if it's really mandatory to have one.

No it is not mandatory to have shared storage. But cluster needs to reach quorum
before it's functional. You need to have more than half of the quorum votes
present or your cluster does not work.

> Maybe i just have to add one vote to the second node ?

Then that node becomes a single point of failure. You need to have 3rd system
for 3rd quorum vote. Either shared storage or quorum server.

Sami
0
11/11/2012 5:22:41 PM
> what does scstat -q say? or clquorum list -v

Here is the result :


-bash-3.2# scstat -q

-- Quorum Summary from latest node reconfiguration --

  Quorum votes possible:      1
  Quorum votes needed:        1
  Quorum votes present:       1


-- Quorum Votes by Node (current status) --

                    Node Name           Present Possible Status
                    ---------           ------- -------- ------
  Node votes:       master              1        1       Online
  Node votes:       slave               0        0       Online


-- Quorum Votes by Device (current status) --

                    Device Name         Present Possible Status
                    -----------         ------- -------- ------
  Device votes:     /dev/did/rdsk/d3s2  0        0       Offline

-bash-3.2# clquorum list -v
Quorum              Type
------              ----
d3                  shared_disk
master              node
slave               node

Impossible to add a vote to SLAVE and D3 .... :o(
0
nslmanu (4)
11/11/2012 6:55:47 PM
On 11/11/2012 4:03 AM, nslmanu wrote:
> Hello,
>
> First : thanks for reading ... :o)
>
> I've just installed a Sun Cluster 3.3 on 2 sol 10 x86 nodes ( 2 VMs on ESXi 5.1 ).
>
> Node Name       Present       Possible       Status
> ---------       -------       --------       ------
> master          1             1              Online
> slave           0             0              Online
>
> All seems to be ok for the moment.
>
> I'm doing a test : Hard Power Off the first node, just to check if the quorum switch to the slave.
>
> But few seconds after this : the slave freeze and reset and stays at boot waiting a for a second node. If i don't boot the master, the slave does not want to boot.
>
> It's is normal ? it's very strange and every time i'm doing this test. Why my second node reset when i power off the master !
>
> Thanks you for help. And sorry for my english. I'm french.
>
> Regards
> Manu
>

This is normal behavior for a cluster.  If a two node cluster member 
loses contact with the other it cannot continue.  Consider the
case of such a cluster when both members want to use the same 
non-shareable resource!  If nothing serializes those transactions
you are looking at a disaster in the making.
0
rgilbert88 (4439)
11/13/2012 12:38:23 AM
Hi guys,

I just finished to add my Quorum device.

=== Cluster Quorum ===
--- Quorum Votes Summary from (latest node reconfiguration) ---
            Needed   Present   Possible
            ------   -------   --------
            2        3         3
--- Quorum Votes by Node (current status) ---
Node Name       Present       Possible       Status
---------       -------       --------       ------
slave           1             1              Online
master          1             1              Online
--- Quorum Votes by Device (current status) ---
Device Name       Present      Possible      Status
-----------       -------      --------      ------
d2                1            1             Online


But i had s very strage behaviour :
Before adding my quorum device, one of my 2 nodes had just 1 vote and the second : 0 vote possible.

And after adding the Quorum device, one vote more appears for the second node !! and of course one vote more from the quorum device.

Why i had a vote added automatically for my second node ? It's very strange , no ?

Thanks for your answer and reading me.

Regards
Manu
0
nslmanu (4)
11/29/2012 2:41:48 PM
Reply:

Similar Artilces:

Sun Cluster 3.0: last quorum lost in not running two node cluster
Hi, where can I find the manual procedure to recover the quorum in case of a not running two node SUN Cluster 3.0 by changing the CCR? Thanks Bernhard ...

Sun Cluster 3.x Jumpstart Nodes
Hello Sun guru's! We are planning to install a Sun Cluster 3.1 with Oracle 9i RAC on Solaris 9 (based on SUNWCXall meta cluster). I am wondering if the following "drivers" of JASS 4.0 are usable for such a configuration: suncluster3x-config.driver suncluster3x-hardening.driver suncluster3x-secure.driver Are there any obvious or less obvious problems I might encounter? Looking forward to receiving your comments I remain, with best regards, Fred ...

Crash-Crash-Crash
Eudora's acting up agian. Latest release, it's been doing this since I paid to upgrade to 6. (nice reward eh?) It runs fine for a few days, then it gets into a fit of crashing. When it crashes, on the next time I run the program, tons of emails that I had already received and filtered are in my inbox, unfiltered, with a few new ones. Sometimes I don't even get done cleaning the mess from the crash before it crashes again. Now I have thousands of duplicate mails in all my mail folders, and no practical way to find and eliminate them.. -- www.MicroBrix.com Little boxes that make life easier. "Dave VanHorn" <dvanhorn@cedar.net> wrote > When it crashes, on the next time I run the program, tons of emails that I > had already received and filtered are in my inbox, unfiltered, with a few > new ones......> Now I have thousands of duplicate mails in all my mail > folders, and no practical way to find and eliminate them.. Question: Do you have Tools > Options > Incoming Mail > "leave mail on server" checked? If so, you keep on picking up the same mail from the server. Something is happening that convinces the server that you haven't picked it up form the same machine before. Is it a laptop that changes connection points? > Question: Do you have Tools > Options > Incoming Mail > "leave mail on > server" checked? Nope. ...

10.3 Crashing And Crashing
http://www.google.com/search?q=10.3+crashing In article <vrhi4c8s9av0f9@news.supernews.com>, "John" <nospam@nospam.com> wrote: > http://www.google.com/search?q=10.3+crashing I'll raise you.. http://www.google.com/search?q=XP+crashing&hl=en&lr=&ie=UTF-8&sa=N&tab=gw In article <vrhi4c8s9av0f9@news.supernews.com>, "John" <nospam@nospam.com> wrote: > http://www.google.com/search?q=10.3+crashing 8,850 hits! I wonder how many hits "Windows XP crashing" yields? 43,500 hits! Hence Windows XP is five times as unstable as Mac OS X 10.3 - right? -- Sandman[.net] Sandman wrote: > In article <vrhi4c8s9av0f9@news.supernews.com>, "John" > <nospam@nospam.com> wrote: > >> http://www.google.com/search?q=10.3+crashing > > 8,850 hits! > > I wonder how many hits "Windows XP crashing" yields? > > 43,500 hits! > > Hence Windows XP is five times as unstable as Mac OS X 10.3 - right? Wrong. Win XP boxes outnumber 10.3 boxes by A LOT MORE THAN 5 to 1. Therefore XP has fewer crash problems per machine than 10.3. "John" <nospam@nospam.com> wrote on 11/17/03 6:09 AM: > http://www.google.com/search?q=10.3+crashing Glad to see you are learning to use Google. See, web searching can be easy. If you need any help, just ask. Snit wrote: > "John" <nospam@nospam.com> wrote on 11/17/03 6:09 AM: >...

How to tell which node is primary in Sun Cluster 3.1
How do I tell which node is primary? We currently have everything running on node A and will be shutting down node B for maintenance, what is to stop resource groups from migrating to node B when it comes up(after hardware maintenance reboot)? On 10/05/12 02:10, Dariusz Dolecki wrote: > How do I tell which node is primary? We currently have everything > running on node A and will be shutting down node B for maintenance, > what is to stop resource groups from migrating to node B when it > comes up(after hardware maintenance reboot)? Run scrgadm -pv and check the failback se...

SN#21710 Sun Blade X6240 with Solaris Cluster 3.1 and 3.2
SYSTEM NEWS FOR SUN USERS Vol 135 Issue 1 2009-05-09 Article 21710 from section "Servers" Allows for Deployment in Cluster Enviornment for Better Service Levels, Higher Availability Solaris Cluster 3.1 and 3.2 now supports the Sun Blade X6240 - a two-socket blade server with Quad-Core AMD Opteron processors and 2:1 (DIMM:core) ratio. Sun Blade X6240 server combined with Solaris Cluster are ideal for mission-critical deployments requiring reliable performance and high service levels. Details at http://sun.systemnews.com/g?A=21710 Have a custom version of 'System News for Sun Users' delivered to you via email each week in text or HTML. Only the sections that you select will be included in your copy of the news magazine. Subscribe at http://sun.systemnews.com/subscribe Free 4-week trial is available. (c) 2008 System News, Inc. http://www.systemnews.com ...

HACMP 5.3 3-nodes cluster, calls network_down 5 sec after node_up by itself
Hi, We do have have problems with an 3-node HACMP cluster ver. 5.3 So the scenario is that we have 3 nodes and a concurent RG and all the 3 nodes are participating in it. Each node has private and public interface (the minimum scenario, no IP address or alias take over defined yet), the network connectivity is just fine, the configuration is double and tripple checked by 3 CEs and big number of references. The verification and synchronization is completing with 'Ok' with no errors and warnings. we have the application server start and stop scripts in place (blank - it's easier to d...

Sun Cluster 3.0 upgrade to Cluster 3.1
Hi How easy is it to upgrade Sun Cluster 3.0 to 3.1. Is there any particular things to look out for in doing the upgrade. My system has the following characteristics. Two nodes Utra-5 and Ultra-10 with 2 Sun 711 Multipacks hosting shared disks, running Solaris 9 using SVM disksets (i.e. not VxVM). The cluster hosts Oracle 9iR2 RAC databases. I'm not sure of the Cluster 3.0 version as it was downloaded from the Sun website. The system uses Sun's global file system. Thanks for any points you can give me. I'm not a cluster expert and this is just my home system for self learning. ...

SN#21521 Solaris Cluster 3.1 and 3.2 Now Support Sun Netra CP3260
SYSTEM NEWS FOR SUN USERS Vol 134 Issue 2 2009-04-11 Article 21521 from section "Telco" Enabling Delivery of Higher Service Levels at Lower Costs and with Reduced Risk Solaris Cluster 3.1 and 3.2 now support the Sun Netra CP3260 ATCA Blade Server - a second-generation telecom blade that meets Advanced TCA standards. Solaris Cluster is a multi-system, multi-site disaster recovery solution and it can enhance the value proposition for the Netra CP3260 servers by enabling delivery of higher service levels, at lower service level costs and reduced risk, Sun reports. Details at http://sun.systemnews.com/g?A=21521 Have a custom version of 'System News for Sun Users' delivered to you via email each week in text or HTML. Only the sections that you select will be included in your copy of the news magazine. Subscribe at http://sun.systemnews.com/subscribe Free 4-week trial is available. (c) 2008 System News, Inc. http://www.systemnews.com ...

SN#21287 Solaris Cluster 3.2 Second Update Release
SYSTEM NEWS FOR SUN USERS Vol 132 Issue 1 2009-02-07 Article 21287 from section "Solaris" Offers Out-of-the Box Support for Largest Number of Commercial and Open-source Apps Solaris Cluster is a multi-system, multi-site disaster recovery solution that manages the availability of applications, services and data across local, regional, and vastly dispersed data centers. This most recent is a freely available download and extends its support for virtualization with the new Solaris Container Cluster, providing business continuity capabilities for deployment of solutions, such as Oracle RAC databases, in Solaris virtualized environments. It also enables increased HA, improved ease of use, flexibility, more application support, and more support for application versions out-of-the-box. Details at http://sun.systemnews.com/g?A=21287 Have a custom version of 'System News for Sun Users' delivered to you via email each week in text or HTML. Only the sections that you select will be included in your copy of the news magazine. Subscribe at http://sun.systemnews.com/subscribe Free 4-week trial is available. (c) 2008 System News, Inc. http://www.systemnews.com ...

MINRES crash on 3 by 3 example!
Hi Let b = [-1;-1;-1] a= [[ 1 0 1];[0 1 1];[1 1 0]] and then do minres(a,b) then I get the output below. That must imply a bug in the minres routine. Or? Erling Warning: Divide by zero. > In minres at 337 Warning: Divide by zero. > In minres at 341 Warning: Divide by zero. > In minres at 339 minres stopped at iteration 3 without converging to the desired tolerance 1e-006 because the maximum number of iterations was reached. The iterate returned (number 0) has relative residual 1 ans = 0 0 0 ...

Eclipse 3.3.1 crash
I copied Eclipse 3.3.1 on my /home/mario directory on Ubuntu 7.04 and Centrino Duo but with only 512MB RAM. I know that it is not enough for any bigger project but at this moment I only learn creating web application. I create project and start help and I my Eclipse simple disapeared from memory. Until that moment it works preety fast. So what could be a problem ? Maybe I need more memory reservating for it, but I don't know how. ...

3.3 box crashes on traffic
Hi NG! My webserver seems to be very unstable since I added new hardware (epox mainboard with nforce2), I assume the nic (D-Link 550TX) using the ste driver is the troublemaker. The box crashes at a high load of ftp and/or http requests. I don't know how to track the crash, maybe it comes to a kernel panic? how can i find out? I will replace the nic as soon as possible, hope it will get better then. kindly, andreas dmesg output: OpenBSD 3.3 (GENERIC) #44: Sat Mar 29 13:22:05 MST 2003 deraadt@i386.openbsd.org:/usr/src/sys/arch/i386/compile/GENERIC cpu0: AMD Athlon XP...

Data Migration on Sun Cluster Nodes (2.2 & 3.1 Versions)
Just wondering if any one had migrated data from an Old array to New array ( We are migrating from an 8830 to DMX )on Sun Cluster nodes running 2.2 (Yes! 2.2...) & Sun Cluster nodes ( 3.1 ). we are moving to a new Fabric as well. My main concern here is with the quorum device. Here is what I have figured out so far. I would appreciate if any one can add more info or any steps I might be missing. Both the nodes can be taken down during this migration. So I dont have to worry about the Reboot as standalone nodes. Present the new Storage to the nodes over the new fabric. Change the quorum ...

CW 8.3 crashing when ANY CW window is resized
This just started happening, after (coincidentally?) installing the 10.3.3 update. Resizing a CW window causes the app to instantly "go away", with the generic error report dialog to Apple. Anyone else seeing this? >This just started happening, after (coincidentally?) installing the 10.3.3 >update. Resizing a CW window causes the app to instantly "go away", with >the generic error report dialog to Apple. Well, not sure why this was happening (my desktop would also crash each time I clicked on the Trash Can) but a reboot fixed it. In article...

SW: Squid cache updated to 3.0.stable26
++ From the VOICE OS/2-eCS News Service http://www.os2voice.org ++ From: madodelDESPAM@DESPAMptdprolog.net Paul Smedley has recompiled Squid cache updated to 3.0.stable26 - with GCC 3.3.5 to prevent a crash. http://os2ports.smedley.info/ http://os2ports.smedley.info/index.php?page=squid "Squid is... - a full-featured Web proxy cache - free, open-source software" Download: http://sourceforge.net/projects/ecsports/files/Squid%20Cache/squid-3.0.stable26-os2-20111220.zip/download Requires: ftp://ftp.netlabs.org/pub/gcc/libc-0.6.4-csd4.zip Please let Paul k...

PHP 5.3.3 Crashes on Win7
Running version 5.3.3 (cli) on Win7/64 runs well from the cmd.exe line after a virgin install. Including ANY DLL in the [PHP] section of the php.ini file even if it doesn't exist (e.g. extension=php_foo.dll) crashes PHP (e.g. php --version) in php5ts.dll as in Problem signature: Problem Event Name: APPCRASH Application Name: php.exe Application Version: 5.3.3.0 Application Timestamp: 4c474a74 Fault Module Name: php5ts.dll Fault Module Version: 5.3.3.0 Fault Module Timestamp: 4c4749e1 Exception Code: c0000005 Exception Offset: 000f5db0 OS Version: 6.1.7600...

CM:SF ... crash ... crash ... crash ... crash ...
Get to map, click on vehicle, watch the game hang for a weirdly long time while the unit data is populated. Then click on any second vehicle and - whack! - desktop city. Every time. That's two in a row for Battlefront. I'm starting to believe that their eLicensing is a moot point, since they don't seem to be able to release anything in a working state. And I'm just loving some of the tech support posts in their forums. Lots of problems with multicore systems and nVidia GPUs - but heck, it's not like 3D game fans gravitate towards dual-core CPUs and...

Trouble compiling qt-x11-free-3.3.3 Solaris 8 SPARC g++ 3.4.3 gmake 3.79.1
QTDIR=/usr/local/src/qt PATH=$QTDIR/bin:$PATH MANPATH=$QTDIR/doc/man:$MANPATH export QTDIR PATH MANPATH cd $QTDIR ../configure -platform solaris-g++ && gmake This results in: [...] g++ -c -Wall -W -O2 -fPIC -DQT_SHARED -DQT_NO_DEBUG -DQT_NO_CUPS -D_LARGEFILE_SOURCE -D_LARGE_FILES -D_FILE_OFFSET_BITS=64 -DQT_NO_XINERAMA -DQT_NO_XCURSOR -DQT_NO_XRANDR -DQT_NO_XRENDER -DQT_NO_XFTFREETYPE -DQT_NO_IMAGEIO_MNG -DQT_NO_IMAGEIO_JPEG -DQT_NO_STYLE_MAC -DQT_NO_STYLE_AQUA -DQT_NO_STYLE_INTERLACE -DQT_NO_STYLE_WINDOWSXP -DQT_NO_STYLE_COMPACT -DQT_NO_STYLE_POCKETPC -I/usr/local/src/qt-x11-free...

kde 3.3.2 crashing while initializing devices
Hi! Since I upgrades to KDE 3.3.2 I can't login to kde anymore. KDE is crashing while initializing devices. Anyone else seen this? Any hints where to look? I've tried to blanken all user dependent stuff. No help. Keeps crashing. I can on the other hand start any kde-application with twm. No problem. Only the kde desktop isn't available. arts-1.3.2,1 Audio system for the KDE integrated X11 desktop de-kde-i18n-3.3.2 German messages and documentation for KDE3 kde-3.3.2 The "meta-port" for KDE kdebase-3.3.2 Basic applications for the KDE system uname -a: FreeBSD hazel 4.11-STABLE FreeBSD 4.11-STABLE #18: Thu Jan 6 09:51:08 CET 2005 HAZEL i386 -- Thomas Thomas Schweikle wrote: > Hi! > > Since I upgrades to KDE 3.3.2 I can't login to kde anymore. KDE is > crashing while initializing devices. Anyone else seen this? Any > hints where to look? > > I've tried to blanken all user dependent stuff. No help. Keeps > crashing. I can on the other hand start any kde-application with > twm. No problem. Only the kde desktop isn't available. > > arts-1.3.2,1 Audio system for the KDE integrated X11 desktop > de-kde-i18n-3.3.2 German messages and documentation for KDE3 > kde-3.3.2 The "meta-port" for KDE > kdebase-3.3.2 Basic applications for the KDE system > > uname -a: > FreeBSD hazel 4.11-STABLE...

Solaris Cluster 3.1 to 3.2 upgrade
On a fully patched Solaris 10 system I'm attempting to upgrade a Solaris Cluster 3.1 to 3.2 using the instructions given in http://docs.sun.com/app/docs/doc/819-2970/gcsrz?a=view using the "standard" upgrade process. All goes well until I attempt to upgrade the data service packages (section 15.) The following question is asked... You must specify the location of the Java Enterprise System (JES) distribution that contains the Sun Cluster data service agents. The name that you give must be the full path to the directory that contains the "Solaris_sparc&...

SN#18280 Configuring Sun Java[TM] System Messaging Server 6.3 with Sun[TM] Cluster 3.1 Software
SYSTEM NEWS FOR SUN USERS Vol 112 Issue 4 2007-06-25 Article 18280 from section "Sun Java System" For Two-Node Asymmetric, HA Cluster An in-depth guide for x86 and SPARC[R] based systems to configuring the Sun Java[TM] System Messaging Server 6.3 with Sun[TM] Cluster 3.1 software is available from the Sun Web site by author Durga Deep Tirunagari. Detailed steps are included in the guide to install and configure (and uninstall) a two-node, asymmetric, high availability (HA) cluster for Sun Java System Messaging Server 6.3 with Sun Cluster 3.1 software. Details at http://sun.systemnews.com/g?A=18280 Have a custom version of 'System News for Sun Users' delivered to you via email each week in PDF, text or HTML. Only the sections that you select will be included in your copy of the news magazine. Subscribe at http://sun.systemnews.com/subscribe (c) 2005 System News, Inc. http://www.systemnews.com ...

gcc 3.3.10 on Power (AIX 3.3.3)
Is there any issue using this compiler on a Power RS/6000 ? Can it compile 64 bit? ...

Sun Cluster 3 vs.Veritas VCS 3.5
I have used Sun Cluster 3 (with VxVM) in the past, but never VCS; can some enlighten individual tell me: The pros and cons of Cluster 3 vs. VCS. The major differences between the two. Which you consider the better product and why? Thanks Fores fores <rjm@fores.plus.com> wrote: > I have used Sun Cluster 3 (with VxVM) in the past, but never VCS; can some > enlighten individual tell me: > > The pros and cons of Cluster 3 vs. VCS. > The major differences between the two. > > Which you consider the better product and why? > VCS is better if you have 3rd part...

Web resources about - Sun Cluster 3.3 : second node crash when first node crash - comp.unix.solaris

Thought Clusters – Software Development and Management
Thought Clusters takes on topics related to the art and science of software development.

Phonological history of English consonant clusters - Wikipedia, the free encyclopedia
The hole–whole merger is the replacement of /hw/ with /h/ before the vowels /oː/ and /uː/ which occurred in Old English . This is due to the ...

No case for dismantling Canberra cluster
&nbsp;Diluting such clusters of key people in research to pork barrel in the bush is risky.

Suicide cluster fears appear to encircle communities at risk engendering panic and fright
Many schools still did not offer suicide prevention programs and he warned that ignoring the issue could be very harmful to students, child psychologist ...

Report: Russia in 'daily cluster bomb attacks' in Syria
... Over the past two weeks, Syrian government and Russian military forces have carried out daily air strikes using internationally banned cluster ...

Book explores "genius clusters" in the world
It's a word that gets thrown around a lot these days, but where exactly does "genius" come from?

Spectacular new Hubble photo shows 'Trumpler 14' star cluster
... a collection of about 2,000 young stars some 8,000 light-years from Earth. This NASA/ESA Hubble Space Telescope image features the star cluster ...

The most massive, distant galaxy cluster: IDCS J1426.5+3508
The most massive, distant galaxy cluster: IDCS J1426.

Japan launches mission to study monster black holes, galaxy clusters
A space observatory that allows scientists to peer into monster-sized black holes and the remains of titanic explosions of stars was successfully ...

Raspberry Pi Zero Cluster Packs a Punch
If you could actually buy 16 Raspberry Pi Zeros, you might be able to build your very own Raspberry Pi Cluster for only $80! Well… minus the ...

Resources last updated: 3/10/2016 8:04:25 PM