GPFS question.

I have some questions regarding the restriping of a GPFS file system (version 3.1.0.9). For the sake of simplicity the cluster is consisted from only 1 node and no customization to the FS, just default settings. The GPFS was created with 2 disks in it, I suspended the second disk, and moved some dat ...

GPFS - mmfsck error

MrsI ran a mmfsck in my GPFS File system . I ´ve received the follow messages.Does anyone have any idea about this message ?Have I to correct my FileSystem ?What is this message , about unreferenced blocks .Is it seriuos this messages ?I ran the mmfsck as "-n" option without correct .Checking "lvdb ...

Problems fixed in GPFS 3.2.1.17

January 14, 2010* Stripe group configuration change so data block loss cannot occur if data is being ingested along with the configuration changes.* Fix policy handling of rules of the form "EXCLUDE FROM POOL" to prevent LOW_SPACE events from incorrectly being logged.* Fix problem on systems configu ...

Problems within optimizing gpfs performance

all GPFS experts!I have been working with gpfs couple of weeks now. We have following hardware setup. Disksystem is EMC CX3-40, gpfs1 & gpfs2 servers are directly connected to emc diskbox via qlogic 2400 fibre card (4gig). gpfs1 & gpfs2 servers are connected to computing cluster via 10gb hp ...

GPFS file system umounted unexpected

All,Our customer has a two node GPFS cluster, running Oracle RAC. They are testdb1i testb2i. The testdb1i is used as nsd server for all nsds.server: p6-570 4core/32G memoryos: AIX5306GPFS:2.3.0.25storage: HDSUser rebooted their system on July 22 and saw one of their gpfs filesystem gsapdata4_lv umou ...

GPFS io performance with different thread combination

I run into a strange problem when I try to evaluate GPFS file system's performance via iozone. Below is my env:8 HS21 nodes as NSD serversAll NSD servers are SAN direct-attached. Total 16 NSDs and each server is used to serve two NSDs.1 HS21 node as NSD cient1 x3950 m2 server as NSD clientInterconne ...

How a file in GPFS is laid out across NSDs

HiIs there a way a find how a 'particular file' (eg. /gpfs1/filename1) is laid out across the LUNs (NSDs) ?Background: I'm testing a gpfs filesystem consisting of 12 LUNs (SAN attached). Each LUN consists of seven 1TB disks in RAID5 (6+1P).I'm interested in finding out how a 10mb file in laid out ac ...

GPFS 3.2.1.10-How to change vgname in lspv output to the correct NSD name?

I have configured a 2-node GPFS 3.2.1.10 cluster. Then the customer wants to reinstall 1 node (node 1), so I mmdelnode it then reinstall AIX on it and install GPFS, then mmaddnode.Now everything seems to be normal except one thing 'lspv does not show the correct NSD name': (I noticed that the vgname ...

gpfs crashed

on one of our computenodes the GPFS crashed and I don't know why.We are running gpfs 3.2.1-8 on sles10 sp1 via ipoib.We see the following logs:node029:/var/adm/ras # cat mmfs.log.previousThu Mar 26 14:05:51 CET 2009 runmmfs startingRemoving old /var/adm/ras/mmfs.log.* files:umount: /home_xyz: device ...

Pagepool - utilization and maximum size (GPFS 3.2.1)

.1. How can we know how much pagepool we are using (the utilization) now?(mmfsadm dump <something>?).2. We set pagepool size to 10GB, but there is a message in mmfs.log that:.Thu Oct 8 09:17:37.391 2009: Pagepool has size 7864320K bytes instead of the requested 10485760K bytes..From GPFS 3.2.1 ...

GPFS : Storage replacement

Question : I need a procedure to move data on GPFSDue to storage migration , we need to migrate GPFS data from one storage box to another one.Doing this migration we (can) also need to reduce the number of lun using bigger lun on new system storage .We are using GPFS 3.1 and 3.2, some (but not all) ...

GPFS will not mount (but shows no errors)

When a machine gets rebooted or gpfs gets mmshutdown'd it is not able to come back up I am running gpfs 3.2.1-1 and upgraded some nodes to 3.2.1-16 but the problem remains. Is there anything I can do to debug the cause. This is most strange as gpfs has run without a problem for over 18 months.mmfslo ...

expand inodes number.

I have a TSM installation, with gpfs filesystem, and since 2 days ago, I cannot write or read anything because filesystem is out of inodes.There are any way to increase the number of inodes?? ...

error: server gpfs for NSD nsd01 couldn't be resolve

I have a couple of nsd servers mounting /dev/dm-* device using multipath:Disk name NSD volume ID Device Node name Remarks nsd00DDN 0A0A00434C728016 /dev/dm-29 gpfs07prv server nodensd00DDN 0A0A00434C728016 /dev/dm-1 gpfs08prv server noderoot@gpfs04prv ~# mmlsnsd -F -mDisk name NSD volume ID Device N ...

Imbalanced GPFS disk usage causes "no space left on device" error

we now did run twice into an issue where processes failed with a "no space left on device" error although a "df" command still showed sufficient amount of space available on the file-system (90% Use).A more details analysis revealed that GPFS had generated a severe imbalance in distributing data amo ...

ccstg3 gpfs filesystem is corrupted

We lost power to one of the disk arrays and now I am unable to remount gpfs. the disks belong to ccstg3 gpfs filesystem currently look like this. #mmlsdisk ccstg3 disk driver sector failure holds holds storage name type size group metadata data status availability pool -------- ------- ----- ccstg3n ...

GPFS 3.4.0.3 AIX : core dump of mmdf command

Since upgrading from 3.4.0.2 to 3.4.0.3 on AIX I get a core dump when I issue a "mmdf myfs" command (myfs is mounted) :<root@qs05753>[/root] 0 # mmdf myfs free KB: 6027-2689 The value for --block-size must be the keyword "auto" or the value must be of the form [n]K, [n]M, [n]G or [n] ...

How can I troubleshoot Windows CIFS issues on GPFS 3.4 nodes?

My environment:Five Windows Server 2008 R2 and four Red Hat Enterprise Linux 5.5 nodes running GPFS 3.4.0.4. There are multiple file systems, but I'd like to focus on one called P3. It has two storage pools (metadataOnly and dataOnly) made up of FC and SATA disks respectively. The LUNs are SAN-attac ...

GPFS filesystem deadlock

I have experienced a deadlock for the only one of two running filesystems on gpfs servers. ls, mmdf and so on was not responded.As I investagted, there were no network or disk problems.Especially, on pgfs07 NSD node, CPU was overloaded(70~80%) by mmfsd, and there was no free mem.So,I attatched the p ...

GPFS unmount due to a single failed drive

I have a gpfs cluster with a replication factor of 2 that unmounteddue to losing a single disk in a single failure group. From myunderstanding this shouldn't happen, so I would like to get somefeedback on the way I am handling disk failures to ensure I am takingthe appropriate steps.I have only seen ...

GPFS problem: system crashed several time a week

I have a AIX 5.3 running with GPFS (I am new on it) and it look likes the system crashed several time and don't know how to fix it, the problem should be around oracle RAC. See the errpt: If someone have on ideaThanks in advance for any help or advise.Cheers LABEL: CORE_DUMPIDENTIFIER: A63BEB70Date/ ...

DM_EVENT_CANCEL by GPFS

Does any one know on which cases DM_EVENT_CANCEL is sent by GPFS? I am getting such events when I am actually expecting DM_EVENT_READ/WRITE on Linux x86_64. Whereas the same DM app on AIX works fine for me. ...

dm_path_to_handle in GPFS

does dm_path_to_handle fail in GPFS-DMAPI if the path contains file from other file system, say JFS2? A per XSDM does it return ENXIO? ...

GPFSF: rename filesystem gpfs need some help please

HAve you ever rename a gpfs filesystem if so are you able to give me the good steps?I am looking on internet for some docs in the same time.Thanks in advance for any help.I'll keep you updated, would be also useful in the future ...

GPFS information can be collected by nmon ?

Environment: HACMP V5.2 + GPFS V2.3 + Oracle 9.2.0.8 RAC When I use nmon in production environment, I/O wait always under 1 (0.1 ~0.6)even disk busy 80~100%, And I find the nmon web page, I found nmon introduction, It does not include GPFS, so nmon can’t collect GPFS I/O Data. It’s true? Or nmon ...

GPFS unsupported in AIX 6.1?

In GPFS page, IBM says that the latest supported AIX version is AIX 5.3. Dose it mean that AIX 6.1 will not work correctly with GPFS? or there is no technical support for it yet? ...

mmcrvsd problem with GPFS on raid0 lun

I had GPFS on Raid5 earlier and we moved to Raid0. I am trying to use one of the luns for this. However I am getting error 6027-1661 with mmcrvsd.root@bull/home/manojcat disk_desc_hdisk2hdisk2:bull.gp.cv.commvault.com:gem.gp.cv.commvault.com:dataAndMetadata::::root@bull/home/manojmmcrnsd -F disk_des ...

How to make GPFS to recognize the change in disk size?

My customer has a GPFS cluster, with a file system called gpfs_fs. That filesystem consists of only 1 NSD disk, called gpfs_nsd (the correspondent hdisk is hdisk3)Now the customer increased the capacity of the hdisk3, and AIX recognizes the change, but when issuing mmdf command the gpfs_nsd still sh ...

Ran out of inodes

my server went down and I realized I am out inodes! 100% full.Im just wondering what my options are short of deleting crap?Im currently hosting a few different of my sites on the same slice, with dev/staging/prod setups for a couple, with git repositories, as well as hosting all my own static files, ...

performance penalty on disk i/o with gpfs?

I am new to GPFS and wonder if there is a general performance penalty for i/o on a gpfs filesystem compared to i/o on a local filesystem. Both being based on the same storage technology (say DS8000). Am I right to expect a somewhat lower i/o bandwidth on gpfs because of the remote mirroring? ...

GPFS: recommended NSD size

I am very new to gpfs, we want to implement a gpfs-2node-cluster (linux) which will CNFS-export a single gpfs-filesystem. We use 2 DS3400 and we wanted to use 10x300GBx4 luns (10 Luns per array; 4 raid5 arrays per DS3400; But it turned out, that DS3400 only supports 32 LUNS to be mappedMy question i ...

GPFS and VIOS vitrual disks

I need help in understanding:1. Can GPFS be implemented using Virtual Disks presented to AIX via VIOS serversIf so where can get a start of documentations? ...

NBU 6.5 support for IBM General Parallel File System (GPFS)?

Anyone has any experience with Netbackup 6.5 + GPFS 3.2??I intend to mount GPFS on one of my server and was wondering if Netbackup has any issue backing up the file system(GPFS)?Do I have to apply any patch/fix in order to make the backup successful?Thanks in advance for any input.Rgds ...

How to Migrate disks of GPFS file system

could you please help me to migrate the disks of a GPFS file system to SAN disks..? ...

GPFS Mount Point Issue

We noticed that on one of my cluster nodes the GPFS filesystem mountpoint is not visible when checked through df command. while GPFS is running we can go in the GPFS file system. when we type command pwd it tells us that we are in GPFS file system but we can,t see files in GPFS file system.when we t ...

AIX GPFS Setup

Guys,I've been going through the interesting redbooks about GPFS, they all provide useful information. However I've got a couple of questions and would love to have a hindsight of someone who has already gone through this type of setup.Summary:My Setup consists of a couple of JS23 Blades within a BC ...

GPFS won't start - complaining about missing SSL library

We have a GPFS cluster that refuses to start. Both nodes are complaining that the SSL library is missing. From mmfslog:Wed Dec 3 10:46:55.795 2008: GPFS: 6027-1544 Sum of prefetchthreads(505) and worker1threads(49) exceeds 550. Reducing them to 501 and 49.Wed Dec 3 10:47:14.817 2008: Failed to open ...

GPFS v3.4 AIX (2 Node - Node and Disk Majority)

I'm planning to setup a 2 node GPFS cluster with node and disk majority.So far, I have done the following: Create GPFS Cluster Add 2nd Node (Both are Active) Now I'm at the step where you create NSD disks, I've been reading a multiple of articles online however can someone guide me how to setup a qu ...

GPFS V3.3 announcements

Watch this thread for announcements on the availability of updates for GPFS v3.3. ...

GPFS V3.4 announcements

Watch this thread for announcements on the availability of updates for GPFS v3.4. ...

GPFS v3.2.1 Announcements

Watch this thread for announcements on the availability of updates for GPFS v3.2.1. ...

GPFS FAQ Announcements

Watch this thread for announcements on the availability of updates to the GPFS FAQ. ...

GPFS in RHEL 5.4

We are eager to upgrade to RHEL 5.4 with kernel version 2.6.18-164.el5 (x86_64). When will RHEL 5.4 x86_64 be supported by GPFS ? ...

GPFS 3.3.0 Installation problems

I'm working with a customer to install GPFS 3.3.0 as a trial and we are running into problems during the install. Here is the error we get after we think we are done with the installation:Package version 3.3.0.0 does not match expected version 3.2.1. Ignoring uploaded package.We have tried to instal ...

GPFS v3.1 Announcements

Watch this thread for announcements on the availability of updates for GPFS v3.1. ...

GPFS 3.3 download

I am new to GPFS, wondering where can I downloaqd the evaluation version of GPFS 3.3 for Windows (or 3.2.1).-Jl ...

Welcome to GPFS - Announce forum

This forum will provide announcements and news on the GPFS product. ...

GPFS information can be collected by nmon ?

Environment: HACMP V5.2 + GPFS V2.3 + Oracle 9.2.0.8 RAC When I use nmon in production environment, I/O wait always under 1 (0.1 ~0.6)even disk busy 80~100%, And I find the nmon web page, I found nmon introduction, It does not include GPFS, so nmon can’t collect GPFS I/O Data. It’s true? Or nmon ...

GPFS and warewulf

Is there support for booting a GPFS storage node via warewulf? How about the same for GPFS clients/compute nodes?Also is there support for SLES10 coming for JS21 bladecenter nodes (ppc64)? Alex Cresswell - Systems AnalystRHCE, CCNA, NCLP, NCLEeXcellence in IS Solutions, Inc. (X-ISS)================= ...

Does gpfs provide any cluster aware file lock api

Does gpfs provide any cluster aware file lock api like fcntl or flock.Does gpfs has a dlm (distributed lock manager)?If so does it use the disk or the network? ...