[Lustre-discuss] Fwd: Annual Lustre Community Survey
Forwarding for Jeff: From: Jeff Denworth jdenwo...@ddn.com Date: Thursday, April 4, 2013 5:54 PM To: lustre-commun...@lists.lustre.org lustre-commun...@lists.lustre.org, lustre-discuss@lists.lustre.org lustre-discuss@lists.lustre.org Subject: Annual Lustre Community Survey All, The past year has seen many significant milestones in Lustre development, community participation and user adoption. Some very exciting work has been done to enable performance at radically new levels of scale; usability administrative features also have been added to enable increased technology adoption. This is a perfect time for us all, as a community, to take stock of where we are and what comes next. Please take 2 minutes to respond to this anonymous 8 question, multiple choice survey. Your answers will help inform the community discussion around the Lustre ecosystem, its evolution and its future. A summary of your responses will be presented live in 2 weeks at the Lustre User Group (LUG’13). YOUR contributions to this survey can help ensure that the survey results reflect the needs, wants and desires of the worldwide community. Thank you in advance for a minute of your time, The DDN Team Survey URL: https://www.surveymonkey.com/s/LustreFS ___ Lustre-discuss mailing list Lustre-discuss@lists.lustre.org http://lists.lustre.org/mailman/listinfo/lustre-discuss
[Lustre-discuss] MD benchmarking with mdtest
A request on a call today prodded me to publish our general-purpose Lustre mdtest benchmark plan, the metadata counterpart to the IOR plan I had published before. Hope you may find it useful. Lustre MD Benchmark Methodology using mdtest http://goo.gl/UBs1p Lustre IO Benchmark Methodology using IOR http://goo.gl/7AWwQ ___ Lustre-discuss mailing list Lustre-discuss@lists.lustre.org http://lists.lustre.org/mailman/listinfo/lustre-discuss
Re: [Lustre-discuss] [wc-discuss] New Test Framework Development - Requirements Capture
Renaming Chris' page to "Test Framework Requiremens", I have added an upper-level pagehttp://wiki.opensfs.org/Improving_the_test_frameworkthat references the three areas of focus we identified on our last phonecall: • Development of a completelynew test frameworkthat will provide the capabilities to test Lustre at exascale. • Incremental benefits byimproving current test framework. • Providing better coverage byimproving individual tests.and links to the requirements page.On Dec 1, 2012, at 9:55 AM, "Gearing, Chris" chris.gear...@intel.com wrote:Hi,During the last meeting we decided to setup a Wiki page to allow everyone to capture their thoughts, requirements and ideas for possible inclusion into the new framework environment.This page is now available on the OpenSFS Wiki and we would welcome your input before the 7th December, please provide input how big or small.http://wiki.opensfs.org/New_test_frameworkMany thanksChris-Intel Corporation (UK) LimitedRegistered No. 1134945 (England)Registered Office: Pipers Way, Swindon SN3 1RJVAT No: 860 2173 47This e-mail and any attachments may contain confidential material forthe sole use of the intended recipient(s). Any review or distributionby others is strictly prohibited. If you are not the intendedrecipient, please contact the sender and delete all copies.attachment: winmail.dat___ Lustre-discuss mailing list Lustre-discuss@lists.lustre.org http://lists.lustre.org/mailman/listinfo/lustre-discuss
Re: [Lustre-discuss] [wc-discuss] Seeking contributors for Lustre User Manual
Would it be easier to move the manual back to a Wiki? The low hassle factor of wikis has always been a draw for contribution. The openSFS site is up and running with MediaWiki now (wiki.opensfs.org). On Nov 9, 2012, at 4:09 PM, Dilger, Andreas andreas.dil...@intel.com wrote: In hopes of improving the quality, coverage, and currency of the Lustre User Manual, I'm putting out a call to the Lustre community for contributors to this important resource. The user manual is especially important for new users to Lustre, but has fallen into some disrepair now that there is no longer a dedicated documentation writer for it. As well, there are many sections of the manual that have become outdated over the years (such as example output, command descriptions, etc) that need to be refreshed. The Lustre User Manual is also a component of Lustre that has a much larger pool of potential contributors than the code itself. If you want to contribute to Lustre, but are not able to contribute with patches to the code, then this is a great opportunity to help out. If you benefit from the open source development of Lustre, then contributing to the manual is a chance to return something back to the community. The Lustre Manual is released under a Creative Commons license, so it is open to all of us to improve. While there is not currently a todo list for the areas of the manual that need updating, looking through open LUDOC tickets is one option: http://bugs.whamcloud.com/secure/QuickSearch.jspa?searchString=LUDOC%20open There are a number of existing documentation tickets for features that are under development for the Lustre 2.4 release (for which we expect to be completed internally), but there are also some tickets from users pointing out errors in the document that need to be fixed. Another way to improve the manual is to simply fetch the manual and read some section at random that you are either interested in, or have some knowledge about, and see if any of the text is confusing, outdated, or incorrect and needs to be updated. The PDF and HTML versions of the current manual are available at: http://wiki.whamcloud.com/display/PUB/Documentation The manual source is hosted in a Git/Gerrit repository in Docbook XML format and can be downloaded at: git clone http://git.whamcloud.com/doc/manual lustre-manual An account in Gerrit is required to download the manual and submit patches. There are some wiki pages that describe the process and packages needed to modify, build, and submit patches to the manual: http://wiki.whamcloud.com/display/PUB/Making+changes+to+the+Lustre+Manual Cheers, Andreas -- Andreas Dilger Lustre Software Architect Intel Corporation ___ Lustre-discuss mailing list Lustre-discuss@lists.lustre.org http://lists.lustre.org/mailman/listinfo/lustre-discuss
Re: [Lustre-discuss] [wc-discuss] Remove a inactive OST
I suspect if you restart the clients the problem will go away. On Oct 30, 2012, at 1:49 AM, Alfonso Pardo alfonso.pa...@ciemat.es wrote: More about my problem: If I run in the client the command lfs df -i. I can see the inactive/removed OST: UUID Inodes IUsed IFree IUse% Mounted on cetafs-MDT_UUID97547059223375132 952095460 2% /mnt/data[MDT:0] cetafs-OST_UUID 1907328018889414 183866 99% /mnt/data[OST:0] cetafs-OST0001_UUID 1907328018889304 183976 99% /mnt/data[OST:1] cetafs-OST0002_UUID 1907328018889353 183927 99% /mnt/data[OST:2] cetafs-OST0003_UUID 1907328018889397 183883 99% /mnt/data[OST:3] cetafs-OST0004_UUID 1907328018889372 183908 99% /mnt/data[OST:4] cetafs-OST0005_UUID 1907328018889440 183840 99% /mnt/data[OST:5] cetafs-OST0006_UUID 1907328018889184 184096 99% /mnt/data[OST:6] cetafs-OST0007_UUID 1907328018889292 183988 99% /mnt/data[OST:7] cetafs-OST0008_UUID 1907328018889134 184146 99% /mnt/data[OST:8] cetafs-OST0009_UUID 1907328018889413 183867 99% /mnt/data[OST:9] cetafs-OST000a_UUID 190732801999 184281 99% /mnt/data[OST:10] cetafs-OST000b_UUID 1907328018889393 183887 99% /mnt/data[OST:11] cetafs-OST000c_UUID 1907328018889290 183990 99% /mnt/data[OST:12] cetafs-OST000d_UUID 1907328018889353 183927 99% /mnt/data[OST:13] cetafs-OST000e_UUID 1907328018889349 183931 99% /mnt/data[OST:14] cetafs-OST000f_UUID 1907328018889357 183923 99% /mnt/data[OST:15] cetafs-OST0010_UUID 1907328018889378 183902 99% /mnt/data[OST:16] cetafs-OST0011_UUID 1907328018889385 183895 99% /mnt/data[OST:17] cetafs-OST0012_UUID 19073280 262901416444266 13% /mnt/data[OST:18] cetafs-OST0013_UUID 19073280 262904516444235 13% /mnt/data[OST:19] OST0014 : Resource temporarily unavailable cetafs-OST0015_UUID 7621120 1494736 6126384 19% /mnt/data[OST:21] cetafs-OST0016_UUID 7621120 1495107 6126013 19% /mnt/data[OST:22] cetafs-OST0017_UUID 7621120 1494952 6126168 19% /mnt/data[OST:23] cetafs-OST0018_UUID 7621120 1494865 6126255 19% /mnt/data[OST:24] filesystem summary:97547059223375132 952095460 2% /mnt/data But, if I list the lustre devices in the client with lctl dl, the inactive/removed OST don't exist: 0 UP mgc MGC192.168.11.9@tcp 7ac5cb56-40f1-7183-672b-8d77a7f42d5d 5 1 UP lov cetafs-clilov-81012394e400 132759a4-add7-eaed-3b81-d27b42f97aef 4 2 UP mdc cetafs-MDT-mdc-81012394e400 132759a4-add7-eaed-3b81-d27b42f97aef 5 3 UP osc cetafs-OST-osc-81012394e400 132759a4-add7-eaed-3b81-d27b42f97aef 5 4 UP osc cetafs-OST0001-osc-81012394e400 132759a4-add7-eaed-3b81-d27b42f97aef 5 5 UP osc cetafs-OST0002-osc-81012394e400 132759a4-add7-eaed-3b81-d27b42f97aef 5 6 UP osc cetafs-OST0003-osc-81012394e400 132759a4-add7-eaed-3b81-d27b42f97aef 5 7 UP osc cetafs-OST0004-osc-81012394e400 132759a4-add7-eaed-3b81-d27b42f97aef 5 8 UP osc cetafs-OST0005-osc-81012394e400 132759a4-add7-eaed-3b81-d27b42f97aef 5 9 UP osc cetafs-OST0006-osc-81012394e400 132759a4-add7-eaed-3b81-d27b42f97aef 5 10 UP osc cetafs-OST0007-osc-81012394e400 132759a4-add7-eaed-3b81-d27b42f97aef 5 11 UP osc cetafs-OST0012-osc-81012394e400 132759a4-add7-eaed-3b81-d27b42f97aef 5 12 UP osc cetafs-OST0013-osc-81012394e400 132759a4-add7-eaed-3b81-d27b42f97aef 5 13 UP osc cetafs-OST0008-osc-81012394e400 132759a4-add7-eaed-3b81-d27b42f97aef 5 14 UP osc cetafs-OST000a-osc-81012394e400 132759a4-add7-eaed-3b81-d27b42f97aef 5 15 UP osc cetafs-OST0009-osc-81012394e400 132759a4-add7-eaed-3b81-d27b42f97aef 5 16 UP osc cetafs-OST000b-osc-81012394e400 132759a4-add7-eaed-3b81-d27b42f97aef 5 17 UP osc cetafs-OST000c-osc-81012394e400 132759a4-add7-eaed-3b81-d27b42f97aef 5 18 UP osc cetafs-OST000d-osc-81012394e400 132759a4-add7-eaed-3b81-d27b42f97aef 5 19 UP osc cetafs-OST000e-osc-81012394e400 132759a4-add7-eaed-3b81-d27b42f97aef 5 20 UP osc cetafs-OST000f-osc-81012394e400 132759a4-add7-eaed-3b81-d27b42f97aef 5 21 UP osc cetafs-OST0010-osc-81012394e400 132759a4-add7-eaed-3b81-d27b42f97aef 5 22 UP osc cetafs-OST0011-osc-81012394e400 132759a4-add7-eaed-3b81-d27b42f97aef 5 23 UP osc cetafs-OST0018-osc-81012394e400 132759a4-add7-eaed-3b81-d27b42f97aef 5 24 UP osc cetafs-OST0015-osc-81012394e400 132759a4-add7-eaed-3b81-d27b42f97aef 5 25 UP osc cetafs-OST0016-osc-81012394e400 132759a4-add7-eaed-3b81-d27b42f97aef 5 26 UP osc cetafs-OST0017-osc-81012394e400 132759a4-add7-eaed-3b81-d27b42f97aef 5 It is
Re: [Lustre-discuss] [wc-discuss] joining two differents mgs in one
Two ways I can think of: 1. mount the MGS disks as ldiskfs and copy the config files from the old MGS to the new, and point all the servers to use the new MGS ipaddr with tunefs 2. writeconf all the servers, setting mgsnode as the new one. This will regenerate all the config files on the new MGS. On Jun 28, 2012, at 9:12 PM, Philippe Weill wrote: Hello for moment and for historic reason we have four lustre filesystem on two different mgs-mds each mgs-mds having two filesystems ( 1.8.7wc1 ) we want to change our mgs-mds infrastructure ( new hardware ) and keeping only one mgs and mds server for all the filesystem I know how to move mdt on new hardware and changing IP but how can we join the two distinct mgs device in one Thanks in advance -- Weill Philippe - Administrateur Systeme et Reseaux CNRS/UPMC/IPSL LATMOS (UMR 8190) ___ Lustre-discuss mailing list Lustre-discuss@lists.lustre.org http://lists.lustre.org/mailman/listinfo/lustre-discuss
Re: [Lustre-discuss] [wc-discuss] Lustre 2.2 production experience
I wasn't complaining, just asking ;) On Jun 14, 2012, at 6:27 PM, Andreas Dilger adil...@whamcloud.com wrote: I think the stability of 2.2.0 is comparable to 2.1.0. One issue is about the number of separate maintenance releases that can be tested. If there are many maintenance releases, then each of those branches would get correspondingly less testing time before release. Secondly, there is a limit on the amount of time that can be spent on porting patches to each maintenance release. This system of maintenance vs. feature releases is similar to what is done for Ubuntu Long Term Stability (LTS) regular releases, and Fedora vs. RHEL. While there is a desire to make each release as reliable as possible, the resources needed to maintain all of the releases for a long time would be very high. Cheers, Andreas On 2012-06-14, at 17:48, Nathan Rutman nathan_rut...@xyratex.com wrote: Is there a belief that Lustre 2.2 is any less stable than Lustre 2.1.0? IOW, are the new features introduced in 2.2 believed to introduce more risk? On Jun 9, 2012, at 3:20 PM, Andreas Dilger wrote: I guess the new Lustre release process is similar to how Ubuntu is released. While we do our best to make each release as stable as possible, there is a different expectation for long-term updates of the feature releases and the maintenance releases. Cheers, Andreas On 2012-06-09, at 16:05, Wojciech Turek wj...@cam.ac.uk wrote: Thanks for a quick reply Andreas. I slightly misunderstood the lustre release process and thought that the next stable/production version is 2.2 I am then interested in the experience of people running Lustre 2.1 Cheers Wojciech On 9 June 2012 21:52, Andreas Dilger adil...@whamcloud.com wrote: I think you'll find that there are not yet (m)any production deployments of 2.2. There are a number of production 2.1 deployments, and this is the current maintenance stream from Whamcloud. Cheers, Andreas On 2012-06-09, at 14:33, Wojciech Turek wj...@cam.ac.uk wrote: I am building a 1.5PB storage system which will employ Lustre as the main file system. The storage system will be extended at the later stage beyond 2PB. I am considering using Lustre 2.2 for production environment. This Lustre storage system will replace our older 300TB system which is currently running Lustre 1.8.8. I am quite happy with lustre 1.8.8 however for the new system Lustre 2.2 seem to be a better match. The storage system will be attached to a university wide cluster (800 nodes), hence there will be quite a large range of applications using the filesystem. Could people with production deployments of Lustre 2.2 share their experience please? -- Wojciech Turek ___ Lustre-discuss mailing list Lustre-discuss@lists.lustre.org http://lists.lustre.org/mailman/listinfo/lustre-discuss ___ Lustre-discuss mailing list Lustre-discuss@lists.lustre.org http://lists.lustre.org/mailman/listinfo/lustre-discuss
Re: [Lustre-discuss] Metadata storage in test script files
On May 4, 2012, at 7:46 AM, Chris Gearing wrote: Hi Roman, I think we may have rat-holed here and perhaps it's worth just re-stating what I'm trying to achieve here. We have a need to be able to test in a more directed and targeted manner, to be able to focus on a unit of code like lnet or an attribute of capability like performance. However since starting work on the Lustre test infrastructure it has become clear to me that knowledge about the capability, functionality and purpose of individual tests is very general and held in the heads of Lustre engineers. Because we are talking about targeting tests we require knowledge about the capability, functionality and purpose of the tests not the outcome of running the tests, or to put it another way what the tests can do not what they have done. One key fact about cataloguing the the capabilities of the tests is that for almost every imaginable case the capability of the test only changes if the test itself changes and so the rate of change of the data in the catalogue is the same and actually much less than the rate of change test code itself. The only exception to this could be that a test suddenly discovers a new bug which has to have a new ticket attached to it, although this should be a very very rare if we manage our development process properly. This requirement leads to the conclusion that we need to catalogue all of the tests within the current test-framework and a catalogue equates to a database, hence we need a database of the capability, functionality and purpose of the individual tests. With this requirement in mind it would be easy to create a database using something like mysql that could be used by applications like the Lustre test system, but using an approach like that would make the database very difficult to share and will be even harder to attach the knowledge to the Lustre tree which is were it belongs. So the question I want to solve is how to catalogue the capabilities of the individual tests in a database, store that data as part of the Lustre source and as a bonus make the data readable and even carefully editable by people as well as machines. Now to focus on the last point I do not think we should constrain ourselves to something that can be read by machine using just bash, we do have access to structure languages and should make use of that fact. I think we all agree 100% on the above... The solution to all of this seemed to be to store the catalogue about the tests as part of the tests themselves ... but not necessarily that conclusion. , this provides for human and machine accessibility, implicit version control and certainty the what ever happens to Lustre source the data goes with it. It is also the case that by keeping the catalogue with the subject the maintenance of the catalogue is more likely to occur than if the two are separate. I agree with all those. But there are some difficulties with this as well: 1. bash isn't a great language to encapsulate this metadata 2. this further locks us in to current test implementation - there's not much possibility to start writing tests in another language if we're parsing through looking for bash-formatted metadata. Sure, multiple parsers could be written... 3. difficulty changing md of groups of tests en-mass - eg. add slow keyword to a set of tests 4. no inheritance of characteristics - each test must explicitly list every piece of md. This not only blows up the amount of md it also is a source for typos, etc. to cause problems. 5. no automatic modification of characteristics. In particular, one piece of md I would like to see is maximum allowed test time for each test. Ideally, this could be measured and adjusted automatically based on historical and ongoing run data. But it would be dangerous to allow automatic modification to the script itself. To address those problems, I think a database-type approach is exactly right, or perhaps a YAML file with hierarchical inheritance. To some degree, this is a evolution vs revolution question, and I prefer to come down on the revolution-enabling design, despite the problems you list. Basically, I believe the separated MD model allows for the replacement of test-framework, and this, to my mind, is the majority driver for adding the MD at all. My original use of the term test metadata is intended as a more modern term for catalogue or the [test] library. So to refresh everybody's mind, I'd like to suggest that we place test metadata in the source code itself using the following format, where the here doc is inserted into the copy about the test function itself. === TEST_METADATA Name: before_upgrade_create_data Summary: Copies lustre source into a node specific directory and then creates a tarball using that directory Description:
Re: [Lustre-discuss] [wc-discuss] Re: Lustre and cross-platform portability
We had a thought that a FUSE-based client based on liblustre might make sense to avoid the NFS/Samba re-export problems (scalability, coherency), at the potential price of some performance. You mentioned on the phone this morning that someone had already done something like this in the past? If we revive this work, presumably we can drop the Mac native client. On the other hand, I agree that the baggage is large, and it will take much more work to see this through, and without a real champion and/or a real community need, I understand the desire to drop it. On Mar 15, 2012, at 11:38 AM, Andreas Dilger wrote: On 2012-03-15, at 8:22 AM, Tyler Hawes wrote: Having a native Windows Mac client is by far our company's #1 most important feature for the future. We have been seriously considering how we could help make this happen, including putting funds and/or developers toward the cause. Tyler, thanks for your feedback. While at Oracle, there was serious effort put toward having a Windows Native Client. Unfortunately, this died along with other Oracle Lustre development projects, and the proprietary code has been sitting unused inside Oracle since then. Similarly, the MacOS native client project started a couple of years ago, but as yet none of the code has been released. While some work was done for a Solaris server, I don't see much hope in that making progress at all. In both cases, I would be supportive of these efforts if there was a likelihood of them bearing fruit any time in the future, but so far I haven't seen any progress in that direction. If Oracle could somehow release the WNC code and/or NRL can overcome the government roadblocks in their path for releasing the MacOS client (MLC?), it would definitely change the direction of my thinking. As it stands now, we have portability code for Windows/Mac without any ability to even build the code, let alone test it, so it is just a burden for Linux. If there was real value being derived from that code (e.g. actual users), then it might be an acceptable burden. I know we're in the minority on this, but I believe it's because we are not using Lustre for HPC. We use it for post-production of TV/film. There are a few others companies in our industry who have started doing this as well. My point is that, while Lustre currently is focused on the HPC crowd who seem not to care about Windows/Mac, Lustre's maturity is giving it the potential to grow into other uses besides HPC. I wouldn't call them general use, but other high-performance uses. In our industry, where there are a lot of Windows Mac workstations that we want to connect to the Lustre storage, the Linux-only client is a major obstacle to that. I'm sure there are other industries that would benefit from this. Yes, the TV/film industry was one of the prime motivators for WNC. There have always been a small number of users from this market, and growth in this industry (and others) is definitely welcomed. If the community wants to keep Lustre strictly HPC focused and discourage other industries from joining in, then abandoning the bridge (albeit the half-built bridge that it is) to Linux/Windows is a good way to do that. If, on the other hand, there is a desire to get some other industries involved, perhaps with more resources and contribution coming from them, then I think it's important to build on the work that has been done. In that regard, upstream Linux kernel inclusion seems like a very low priority to me. I think even for upstream kernel submission, if it were clear that the layering in Lustre was for a valid purpose (i.e. existing Mac/Win clients) then it might be accepted. As it stands now, we can't honestly make that argument to the kernel maintainers. 2012/3/15 lustre-discuss-requ...@lists.lustre.org -- Forwarded message -- From: Andreas Dilger adil...@whamcloud.com To: t...@lists.opensfs.org Cc: wc-discuss wc-disc...@whamcloud.com, lustre-discuss discuss lustre-discuss@lists.lustre.org, Lustre Devel lustre-de...@lists.lustre.org Date: Wed, 14 Mar 2012 18:31:29 -0600 Subject: [Lustre-discuss] Lustre and cross-platform portability Whamcloud and EMC are jointly investigating how to be able to contribute the Lustre client code into the upstream Linux kernel. As a prerequisite to this, EMC is working to clean up the Lustre client code to better match the kernel coding style, and one of the anticipated major obstacles to upstream kernel submission is the heavy use of code abstraction via libcfs for portability to other operating systems (most notably MacOS and WinNT, but also for liblustre, and potentially *BSD). I have no information that the WinNT project will ever be released by Oracle, and as yet there has not been any code released from the MacOS port, so the libcfs portability layer is potentially exacting a high cost in code
Re: [Lustre-discuss] [Lustre-devel] [wc-discuss] Re: [Twg] Lustre and cross-platform portability
On Mar 16, 2012, at 8:06 AM, Todd, Allen wrote: On Friday, March 16, 2012 6:11 AM, Gregory Matthews wrote: why bother having a windows client if you lose the performance? We have windows based detectors and proprietary windows based analysis software that would definitely benefit from higher performance access to lustre file systems but replacing existing CIFS servers for no gain seems a bit pointless. 1. Re-exporting Lustre via CIFS or NFS isn't scalable to very large numbers of Windows clients 2. Re-exporting Lustre via CIFS or NFS can have coherency problems when multiple re-exporters are involved 3. Wide-striped file access would likely have greater performance on a native client than via a re-exported single pipe, for some value of wide. From the perspective of my firm, the benefit of a windows lustre client (native or fuse-based), even one that performs at 10% or 20% of the linux client, is the improved scalability that it offers. Our current solution uses 4 samba gateways per hundred windows servers to achieve acceptable bandwidth, but that bandwidth can be cut to an unacceptable trickle if a large wave of native linux clients simultaneously accesses the filesystem. We have repeatedly asked our Microsoft HPC contacts to intervene to either fund a native windows client or to get oracle to release the existing one, since scalable storage is a big hole in the Microsoft HPC toolkit. Obviously, nothing has come out of that. It seems Microsoft is banking on pNFS, eventually working in this space. Allen Todd IMPORTANT: The information contained in this email and/or its attachments is confidential. If you are not the intended recipient, please notify the sender immediately by reply and immediately delete this message and all its attachments. Any review, use, reproduction, disclosure or dissemination of this message or any attachment by an unintended recipient is strictly prohibited. Neither this message nor any attachment is intended as or should be construed as an offer, solicitation or recommendation to buy or sell any security or other financial instrument. Neither the sender, his or her employer nor any of their respective affiliates makes any warranties as to the completeness or accuracy of any of the information contained herein or that this message or any of its attachments is free of viruses. ___ Lustre-devel mailing list lustre-de...@lists.lustre.org http://lists.lustre.org/mailman/listinfo/lustre-devel ___ Lustre-discuss mailing list Lustre-discuss@lists.lustre.org http://lists.lustre.org/mailman/listinfo/lustre-discuss
[Lustre-discuss] Test restructuring
Hi Chris - talking to Eric after EOFS he mentioned that you were not just interested in automating the existing Lustre tests but also in restructuring how they were designed / set up to make them more reliable, repeatable, and easier to run and automate. Or, at least that is my desire and I was hoping you had similar thoughts / plans. In particular, I'm interested in a few things: 1. Separating out the setup of the filesystem from the execution of the tests. 2. Reducing the ordering dependencies on the tests and clarifying their individual prerequisites and postconditions. 3. Reorganizing the tests into more logical groupings. 4. Cutting out redundant / ineffective tests. Are you / WhamCloud pursuing any of this at the moment? __ This email may contain privileged or confidential information, which should only be used for the purpose for which it was sent by Xyratex. No further rights or licenses are granted to use such information. If you are not the intended recipient of this message, please notify the sender by return and delete it. You may not use, copy, disclose or rely on the information contained in it. Internet email is susceptible to data corruption, interception and unauthorised amendment for which Xyratex does not accept liability. While we have taken reasonable precautions to ensure that this email is free of viruses, Xyratex does not accept liability for the presence of any computer viruses in this email, nor for any losses caused as a result of viruses. Xyratex Technology Limited (03134912), Registered in England Wales, Registered Office, Langstone Road, Havant, Hampshire, PO9 1SA. The Xyratex group of companies also includes, Xyratex Ltd, registered in Bermuda, Xyratex International Inc, registered in California, Xyratex (Malaysia) Sdn Bhd registered in Malaysia, Xyratex Technology (Wuxi) Co Ltd registered in The People's Republic of China and Xyratex Japan Limited registered in Japan. __ ___ Lustre-discuss mailing list Lustre-discuss@lists.lustre.org http://lists.lustre.org/mailman/listinfo/lustre-discuss
Re: [Lustre-discuss] [Lustre-devel] Queries regarding Lustre Throughput Numbers with mdtest benchmark
I suspect that you are running this test against local hard drives instead of a shared Lustre mount point. Are you sure Lustre is mounted at /tmp/l66 on all clients in hostfile? On Jul 6, 2011, at 7:25 PM, Andreas Dilger wrote: On 2011-06-22, at 5:06 PM, vilobh meshram wrote: I have a query regarding Lustre Throughput Numbers with mdtest benchmark.I am running mdtest benhmark with following options :- /home/meshram/mpich2-new/mpich2-1.4/mpich2-install/bin/mpirun -np 256 -hostfile ./hostfile ./mdtest -z 3 -b 10 -I 5 -v -d /tmp/l66 where , mdtest - is the standard benchmark to test metadata operations. [https://computing.llnl.gov/?set=codepage=sio_downloads ] /tmp/l66 is my Lustre mount. I am using 1Gige Network with TCP transport. hostfile has 8 host nodes I am varying the number of processes as we can see in following table .. I was amazed by the throughput which I got. I think this is too huge. Can someone please let me know if these numbers are correct ? 4 Process : File creation : 0.501 sec, 44392.140 ops/sec 8 Process : File creation : 0.685 sec, 64890.598 ops/sec 16 Process: File creation : 1.426 sec, 62318.798 ops/sec 32 Process: File creation : 2.947 sec, 60312.766 ops/sec 64 Process: File creation : 5.630 sec, 63142.760 ops/sec 128 Process : File creation : 13.208 sec, 53835.707 ops/sec 256 Process : File creation : 24.601 sec, 57804.777 ops/sec Seems nobody has responded to your email, and I just found it buried in my inbox. I agree that the numbers are quite high, especially for GigE networking. More typical numbers are in the 5-20k creates/sec (depending on network and MDS hardware). That said, the above may not be completely impossible for small file counts (allowing all of the creates to be served from cache), or if the client+MDS+OSS are all on the same node (which avoids any network latency). You also didn't describe the Lustre filesystem, nor what version you are testing. This workload has gotten a faster with Lustre 2.1, and it stands to get faster in Lustre 2.2 also. Cheers, Andreas -- Andreas Dilger Principal Engineer Whamcloud, Inc. ___ Lustre-devel mailing list lustre-de...@lists.lustre.org http://lists.lustre.org/mailman/listinfo/lustre-devel __ This email may contain privileged or confidential information, which should only be used for the purpose for which it was sent by Xyratex. No further rights or licenses are granted to use such information. If you are not the intended recipient of this message, please notify the sender by return and delete it. You may not use, copy, disclose or rely on the information contained in it. Internet email is susceptible to data corruption, interception and unauthorised amendment for which Xyratex does not accept liability. While we have taken reasonable precautions to ensure that this email is free of viruses, Xyratex does not accept liability for the presence of any computer viruses in this email, nor for any losses caused as a result of viruses. Xyratex Technology Limited (03134912), Registered in England Wales, Registered Office, Langstone Road, Havant, Hampshire, PO9 1SA. The Xyratex group of companies also includes, Xyratex Ltd, registered in Bermuda, Xyratex International Inc, registered in California, Xyratex (Malaysia) Sdn Bhd registered in Malaysia, Xyratex Technology (Wuxi) Co Ltd registered in The People's Republic of China and Xyratex Japan Limited registered in Japan. __ ___ Lustre-discuss mailing list Lustre-discuss@lists.lustre.org http://lists.lustre.org/mailman/listinfo/lustre-discuss
Re: [Lustre-discuss] Too many client eviction
On May 3, 2011, at 10:09 AM, DEGREMONT Aurelien wrote: Correct me if I'm wrong, but when I'm looking at Lustre manual, it said that client is adapting its timeout, but not the server. I'm understood that server-client RPC still use the old mechanism, especially for our case where it seems server is revoking a client lock (ldlm_timeout is used for that?) and client did not respond. Server and client cooperate together for the adaptive timeouts. I don't remember which bug the ORNL settings were in, maybe 14071, bugzilla's not responding at the moment. But a big question here is why 25315 seconds for a callback - that's well beyond anything at_max should allow... I forgot to say that we have LNET routers also involved for some cases. Thank you Aurélien Andreas Dilger a écrit : I don't think ldlm_timeout and obd_timeout have much effect when AT is enabled. I believe that LLNL has some adjusted tunables for AT that might help for you (increased at_min, etc). Hopefully Chris or someone at LLNL can comment. I think they were also documented in bugzilla, though I don't know the bug number. Cheers, Andreas On 2011-05-03, at 6:59 AM, DEGREMONT Aurelien aurelien.degrem...@cea.fr wrote: Hello We often see some of our Lustre clients being evicted abusively (clients seem healthy). The pattern is always the same: All of this on Lustre 2.0, with adaptative timeout enabled 1 - A server complains about a client : ### lock callback timer expired... after 25315s... (nothing on client) (few seconds later) 2 - The client receives -107 to a obd_ping for this target (server says @@@processing error 107) 3 - Client realize its connection was lost. Client notices it was evicted. It reconnects. (To be sure) When client is evicted, all undergoing I/O are lost, no recovery will be done for that? We are thinking to increase timeout to give more time to clients to answer the ldlm revocation. (maybe it is just too loaded) - Is ldlm_timeout enough to do so? - Do we need to also change obd_timeout in accordance? Is there a risk to trigger new timeouts if we just change ldlm_timeout (cascading timeout). Any feedback in this area is welcomed. Thank you Aurélien Degrémont ___ Lustre-discuss mailing list Lustre-discuss@lists.lustre.org http://lists.lustre.org/mailman/listinfo/lustre-discuss ___ Lustre-discuss mailing list Lustre-discuss@lists.lustre.org http://lists.lustre.org/mailman/listinfo/lustre-discuss ___ Lustre-discuss mailing list Lustre-discuss@lists.lustre.org http://lists.lustre.org/mailman/listinfo/lustre-discuss
Re: [Lustre-discuss] [Lustre-devel] Research on filesystem metadata operation distribution
Hmm - also 'uptime' would really be nice to have with these so we could estimate rough rates... On Apr 21, 2011, at 11:40 AM, Andreas Dilger wrote: I'm trying to get some data about the relative distribution of MDS operations in the wild, and I'd be grateful if some people with production filesystems that have been running for at least a week could collect some simple stats and email them to me. They can be collected by any regular user on the MDS node: lctl get_param mds.*.stats | egrep open|close|rename|link|attr|sync It would be useful to also include lfs df and lfs df -i information, as well as a brief description of what the filesystem is used for (scratch, home, project, archive, etc). As a reminder, I'm also interested if some Lustre admins could run the fsstats tool from http://www.pdsi-scidac.org/fsstats/ and send me the output. Sending the output to PDSI via their submission form may also produce some positive results. http://www.pdsi-scidac.org/fsstats/files/fsstats-1.4.5.tar.gz Thanks in advance for any data. I've set replies to go only to lustre-devel, to avoid clogging the larger readership of lustre-discuss, but it may be useful for others to have this in a list archive and/or searchable via Google in the future so I don't necessarily want to keep it all to myself. Cheers, Andreas -- Andreas Dilger Principal Engineer Whamcloud, Inc. ___ Lustre-devel mailing list lustre-de...@lists.lustre.org http://lists.lustre.org/mailman/listinfo/lustre-devel __ This email may contain privileged or confidential information, which should only be used for the purpose for which it was sent by Xyratex. No further rights or licenses are granted to use such information. If you are not the intended recipient of this message, please notify the sender by return and delete it. You may not use, copy, disclose or rely on the information contained in it. Internet email is susceptible to data corruption, interception and unauthorised amendment for which Xyratex does not accept liability. While we have taken reasonable precautions to ensure that this email is free of viruses, Xyratex does not accept liability for the presence of any computer viruses in this email, nor for any losses caused as a result of viruses. Xyratex Technology Limited (03134912), Registered in England Wales, Registered Office, Langstone Road, Havant, Hampshire, PO9 1SA. The Xyratex group of companies also includes, Xyratex Ltd, registered in Bermuda, Xyratex International Inc, registered in California, Xyratex (Malaysia) Sdn Bhd registered in Malaysia, Xyratex Technology (Wuxi) Co Ltd registered in The People's Republic of China and Xyratex Japan Limited registered in Japan. __ ___ Lustre-discuss mailing list Lustre-discuss@lists.lustre.org http://lists.lustre.org/mailman/listinfo/lustre-discuss
Re: [Lustre-discuss] Lustre community build server
Hi Aurelien, Robert - We also use Hudson and are interested in using it to do Lustre builds and testing. Hi Robert Read a écrit : Hi Aurélien, Yes, we've noticed Hudson's support for testing is not quite what we need, so we're planning to use Hudson to trigger our testing system, but not necessarily to manage it. We'd definitely be interested in learning more about your experiences, though. I do not know what you mean by triggering your testing system. But here is what I set up. Hudson has only 1 slave node dedicated to testing Lustre 2. Hudson will launch a shell script through ssh to it. This script: - retrieves Lustre source (managed by Hudson git plugin) - compiles it. - launches acceptance-small with several parameters. - acceptance-small will connect to other nodes dedicated for these tests. acc-sm have been patched: - to be more error resilient (does not stop at first failure) - to generate a test report in JUNIT format. Is this the yaml acc-sm that Robert was talking about, or an older one? Hudson fetch the junit report and parse it thanks to its plugin. Hudson can display in its interface all tests successes and failures. Everything goes fine as long as: - the testsuite leaves the node in a good shape. It is difficult to have a automatic way to put the node back. Currently, we need to manualy fix that. Would it be helpful to run the test in a VM? Hudson has a libvirt-slave plugin that seems like it can start and stop a VM for you. Another point I like about VM's is that they can be suspended and shipped to an investigator for local debugging. - Hudson does not know about the other nodes used by acc-sm. And so can trigger tests even if some sattelites nodes are unavailable. Don't know if libvirt-slave can handle multiple VM's for a multi-node Lustre test, but maybe it can be extended. How is you do this on your side? It seems that like you, we are more interested in reporting the results within Hudson as opposed to a different custom tool. Nathan Aurélien robert On Dec 16, 2010, at 1:22 , DEGREMONT Aurelien wrote: Hi Robert, That's very interesting. At CEA we also have a Hudson platform and I'm running acceptance-small for several Lustre branches in it. Hudson is a great tool but it was not design to test tools that run kernel-space that can crash your nodes or, at least, put your kernel in a bad shape. I will be very interested to share Hudson experience testing Lustre and see how you've configured it for your own tests. Aurélien Robert Read a écrit : Hi, As I mentioned the other day, Whamcloud is hosting a Hudson build server and producing snapshot builds for Centos 5.x (and Ubuntu 10.4 when it works) for both 1.8.x and 2.x branches. Our intention is for this to be a resource for the Lustre community to find recent Lustre packages for variety of Linux distributions. Early next year we'll connect this to our test system so at least some of the packages can be tested, as well. We would be interested in hearing from anyone that would like to participate producing builds. Hudson is an distributed system, and it's easy to add more build nodes, even behind firewalls (some of us are running build VMs on our home machines). If you would like add another distribution or architecture we don't have yet, or even one we do have (the more the merrier), we'd be happy to work with you to do that. Please contact me if you are interested. cheers, robert ___ Lustre-discuss mailing list Lustre-discuss@lists.lustre.org http://lists.lustre.org/mailman/listinfo/lustre-discuss ___ Lustre-discuss mailing list Lustre-discuss@lists.lustre.org http://lists.lustre.org/mailman/listinfo/lustre-discuss
[Lustre-discuss] Fwd: Lustre and Large Pages
Jim, I'm forwarding this to lustre-discuss to get a broader community input. I'm sure somebody has some experience with this. Begin forwarded message: I am looking for information on how Lustre assigns and holds pages on client nodes across jobs. The motivation is that we want to make huge pages available to users. We have found that it is almost impossible to allocate very many huge pages since Lustre holds scattered small pages across jobs. In fact, typically about 1/3 of compute node memory can be allocated as huge pages. We have done quite a lot of performance studies which show that a substantial percentage of jobs on Ranger have TLB misses as a major performance bottleneck. We estimate we might recover as much as an additional 5%-10% throughput if users could use huge pages. Therefore we would like to find a way to minimize the client memory which Lustre holds across jobs. Have you had anyone else mention this situation to you? Regards, Jim Browne ___ Lustre-discuss mailing list Lustre-discuss@lists.lustre.org http://lists.lustre.org/mailman/listinfo/lustre-discuss
Re: [Lustre-discuss] Problem with write_conf
There's a 'failsafe' feature that prevents filesystem name changes: LustreError: 157-3: Trying to start OBD AFTER-MDT_UUID using the wrong disk BEFORE-MDT_UUID. Were the /dev/ assignments rearranged? You'll have to go and delete the last_rcvd file off the disk for all the servers in the filesystem as well as tunefs --writeconf them all to the name AFTER name. On Aug 2, 2010, at 6:08 PM, Roger Spellman wrote: Hi, I would like to be able to change a file system name. Towards that end, I have run the following commands as an experiment: mkfs.lustre --reformat --fsname BEFORE --device-size=1 --mgs --mdt --mgsnode=10.2@o2ib0 /dev/mapper/map0 dmesg -c mount -t lustre /dev/mapper/map0 /mnt/mdt dmesg -c umount /mnt/mdt dmesg -c tunefs.lustre --writeconf --fsname=AFTER --mgs --mdt /dev/mapper/map0 dmesg -c mount -t lustre /dev/mapper/map0 /mnt/mdt dmesg -c Unfortunately, this does not work. Can someone please explain the correct sequence of commands to ues? The output of each command is as follows. Thanks. [r...@ts-hss2-01 ~]# mkfs.lustre --reformat --fsname BEFORE --device-size=1 --mgs --mdt --mgsnode=10.2@o2ib0 /dev/mapper/map0 Permanent disk data: Target: BEFORE-MDT Index: unassigned Lustre FS: BEFORE Mount type: ldiskfs Flags: 0x75 (MDT MGS needs_index first_time update ) Persistent mount opts: iopen_nopriv,user_xattr,errors=remount-ro Parameters: mgsnode=10.2@o2ib mdt.group_upcall=/usr/sbin/l_getgroups device size = 1632256MB 2 6 18 formatting backing filesystem ldiskfs on /dev/mapper/map0 target name BEFORE-MDT 4k blocks 2500 options-i 4096 -I 512 -q -O dir_index,extents,uninit_groups -F mkfs_cmd = mke2fs -j -b 4096 -L BEFORE-MDT -i 4096 -I 512 -q -O dir_index,extents,uninit_groups -F /dev/mapper/map0 2500 Writing CONFIGS/mountdata [r...@ts-hss2-01 ~]# dmesg -c LDISKFS-fs: barriers enabled kjournald2 starting: pid 1388, dev dm-4:8, commit interval 5 seconds LDISKFS FS on dm-4, internal journal on dm-4:8 LDISKFS-fs: delayed allocation enabled LDISKFS-fs: file extents enabled LDISKFS-fs: mballoc enabled LDISKFS-fs: mounted filesystem dm-4 with ordered data mode LDISKFS-fs: mballoc: 1 blocks 1 reqs (0 success) LDISKFS-fs: mballoc: 1 extents scanned, 0 goal hits, 1 2^N hits, 0 breaks, 0 lost LDISKFS-fs: mballoc: 1 generated and it took 2142 LDISKFS-fs: mballoc: 512 preallocated, 0 discarded [r...@ts-hss2-01 ~]# mount -t lustre /dev/mapper/map0 /mnt/mdt [r...@ts-hss2-01 ~]# dmesg -c LDISKFS-fs: barriers enabled kjournald2 starting: pid 1406, dev dm-4:8, commit interval 5 seconds LDISKFS FS on dm-4, internal journal on dm-4:8 LDISKFS-fs: delayed allocation enabled LDISKFS-fs: file extents enabled LDISKFS-fs: mballoc enabled LDISKFS-fs: mounted filesystem dm-4 with ordered data mode LDISKFS-fs: mballoc: 0 blocks 0 reqs (0 success) LDISKFS-fs: mballoc: 0 extents scanned, 0 goal hits, 0 2^N hits, 0 breaks, 0 lost LDISKFS-fs: mballoc: 0 generated and it took 0 LDISKFS-fs: mballoc: 0 preallocated, 0 discarded LDISKFS-fs: barriers enabled kjournald2 starting: pid 1410, dev dm-4:8, commit interval 5 seconds LDISKFS FS on dm-4, internal journal on dm-4:8 LDISKFS-fs: delayed allocation enabled LDISKFS-fs: file extents enabled LDISKFS-fs: mballoc enabled LDISKFS-fs: mounted filesystem dm-4 with ordered data mode Lustre: MGS MGS started Lustre: mgc10.2@o2ib: Reactivating import Lustre: Setting parameter BEFORE-MDT.mdt.group_upcall in log BEFORE-MDT Lustre: Enabling user_xattr Lustre: BEFORE-MDT: new disk, initializing Lustre: BEFORE-MDT: Now serving BEFORE-MDT on /dev/mapper/map0 with recovery enabled Lustre: 1503:0:(lproc_mds.c:271:lprocfs_wr_group_upcall()) BEFORE-MDT: group upcall set to /usr/sbin/l_getgroups Lustre: BEFORE-MDT.mdt: set parameter group_upcall=/usr/sbin/l_getgroups [r...@ts-hss2-01 ~]# umount /mnt/mdt [r...@ts-hss2-01 ~]# dmesg -c Lustre: Failing over BEFORE-MDT Lustre: Skipped 1 previous similar message Lustre: *** setting obd BEFORE-MDT device 'dm-4' read-only *** Turning device dm-4 (0xfd4) read-only Lustre: BEFORE-MDT: shutting down for failover; client state will be preserved. Lustre: MDT BEFORE-MDT has stopped. LustreError: 1517:0:(ldlm_request.c:1025:ldlm_cli_cancel_req()) Got rc -108 from cancel RPC: canceling anyway LustreError: 1517:0:(ldlm_request.c:1587:ldlm_cli_cancel_list()) ldlm_cli_cancel_list: -108 Lustre: MGS has stopped. LDISKFS-fs: mballoc: 3 blocks 3 reqs (0 success) LDISKFS-fs: mballoc: 8 extents scanned, 0 goal hits, 0 2^N hits, 0 breaks, 0 lost LDISKFS-fs: mballoc: 1 generated and it took 2598 LDISKFS-fs: mballoc: 1145 preallocated, 0 discarded Removing read-only on unknown block (0xfd4) Lustre: server umount BEFORE-MDT complete
Re: [Lustre-discuss] Problem with write_conf
On Aug 3, 2010, at 11:25 AM, Roger Spellman wrote: Nathan, Thank you. That works! I found that if I change IP address, I also need to remove the file /mnt/mdt/CONFIGS/*-client. This is what tunefs.lustre --writeconf on the MDT does, when you first mount it after the writeconf. --writeconf on the MDT and all OSTs is the preferred way of changing a server nid. The reason is that the OST mounts failed – the OST was still looking for the old IP Address. I grepped for files with the old IP Address, and I found those client files. Is that a safe thing to do? Please note that my mdt and mgs are on the same LUN. Thanks. -Roger From: Nathan Rutman [mailto:nathan.rut...@oracle.com] Sent: Tuesday, August 03, 2010 2:03 PM To: Roger Spellman Cc: lustre-discuss@lists.lustre.org Subject: Re: [Lustre-discuss] Problem with write_conf There's a 'failsafe' feature that prevents filesystem name changes: LustreError: 157-3: Trying to start OBD AFTER-MDT_UUID using the wrong disk BEFORE-MDT_UUID. Were the /dev/ assignments rearranged? You'll have to go and delete the last_rcvd file off the disk for all the servers in the filesystem as well as tunefs --writeconf them all to the name AFTER name. On Aug 2, 2010, at 6:08 PM, Roger Spellman wrote: Hi, I would like to be able to change a file system name. Towards that end, I have run the following commands as an experiment: mkfs.lustre --reformat --fsname BEFORE --device-size=1 --mgs --mdt --mgsnode=10.2@o2ib0 /dev/mapper/map0 dmesg -c mount -t lustre /dev/mapper/map0 /mnt/mdt dmesg -c umount /mnt/mdt dmesg -c tunefs.lustre --writeconf --fsname=AFTER --mgs --mdt /dev/mapper/map0 dmesg -c mount -t lustre /dev/mapper/map0 /mnt/mdt dmesg -c Unfortunately, this does not work. Can someone please explain the correct sequence of commands to ues? The output of each command is as follows. Thanks. [r...@ts-hss2-01 ~]# mkfs.lustre --reformat --fsname BEFORE --device-size=1 --mgs --mdt --mgsnode=10.2@o2ib0 /dev/mapper/map0 Permanent disk data: Target: BEFORE-MDT Index: unassigned Lustre FS: BEFORE Mount type: ldiskfs Flags: 0x75 (MDT MGS needs_index first_time update ) Persistent mount opts: iopen_nopriv,user_xattr,errors=remount-ro Parameters: mgsnode=10.2@o2ib mdt.group_upcall=/usr/sbin/l_getgroups device size = 1632256MB 2 6 18 formatting backing filesystem ldiskfs on /dev/mapper/map0 target name BEFORE-MDT 4k blocks 2500 options-i 4096 -I 512 -q -O dir_index,extents,uninit_groups -F mkfs_cmd = mke2fs -j -b 4096 -L BEFORE-MDT -i 4096 -I 512 -q -O dir_index,extents,uninit_groups -F /dev/mapper/map0 2500 Writing CONFIGS/mountdata [r...@ts-hss2-01 ~]# dmesg -c LDISKFS-fs: barriers enabled kjournald2 starting: pid 1388, dev dm-4:8, commit interval 5 seconds LDISKFS FS on dm-4, internal journal on dm-4:8 LDISKFS-fs: delayed allocation enabled LDISKFS-fs: file extents enabled LDISKFS-fs: mballoc enabled LDISKFS-fs: mounted filesystem dm-4 with ordered data mode LDISKFS-fs: mballoc: 1 blocks 1 reqs (0 success) LDISKFS-fs: mballoc: 1 extents scanned, 0 goal hits, 1 2^N hits, 0 breaks, 0 lost LDISKFS-fs: mballoc: 1 generated and it took 2142 LDISKFS-fs: mballoc: 512 preallocated, 0 discarded [r...@ts-hss2-01 ~]# mount -t lustre /dev/mapper/map0 /mnt/mdt [r...@ts-hss2-01 ~]# dmesg -c LDISKFS-fs: barriers enabled kjournald2 starting: pid 1406, dev dm-4:8, commit interval 5 seconds LDISKFS FS on dm-4, internal journal on dm-4:8 LDISKFS-fs: delayed allocation enabled LDISKFS-fs: file extents enabled LDISKFS-fs: mballoc enabled LDISKFS-fs: mounted filesystem dm-4 with ordered data mode LDISKFS-fs: mballoc: 0 blocks 0 reqs (0 success) LDISKFS-fs: mballoc: 0 extents scanned, 0 goal hits, 0 2^N hits, 0 breaks, 0 lost LDISKFS-fs: mballoc: 0 generated and it took 0 LDISKFS-fs: mballoc: 0 preallocated, 0 discarded LDISKFS-fs: barriers enabled kjournald2 starting: pid 1410, dev dm-4:8, commit interval 5 seconds LDISKFS FS on dm-4, internal journal on dm-4:8 LDISKFS-fs: delayed allocation enabled LDISKFS-fs: file extents enabled LDISKFS-fs: mballoc enabled LDISKFS-fs: mounted filesystem dm-4 with ordered data mode Lustre: MGS MGS started Lustre: mgc10.2@o2ib: Reactivating import Lustre: Setting parameter BEFORE-MDT.mdt.group_upcall in log BEFORE-MDT Lustre: Enabling user_xattr Lustre: BEFORE-MDT: new disk, initializing Lustre: BEFORE-MDT: Now serving BEFORE-MDT on /dev/mapper/map0 with recovery enabled Lustre: 1503:0:(lproc_mds.c:271:lprocfs_wr_group_upcall()) BEFORE-MDT: group upcall set to /usr/sbin/l_getgroups Lustre: BEFORE-MDT.mdt: set parameter group_upcall=/usr/sbin/l_getgroups [r...@ts-hss2-01 ~]# umount /mnt/mdt [r...@ts-hss2-01 ~]# dmesg -c
Re: [Lustre-discuss] OST 1.6.6-1.8.3 upgrade
On Jul 27, 2010, at 5:02 AM, Heiko Schröter wrote: Hello, running tunefs.lustre during the upgrade, do i have to add the --param option when it has been used during the creation fo the fs ? i.e. lustre fs during creation: mkfs.lustre --param=failover.mode=failout --fsname lustre --ost --mkfsoptions='-i 2097152 -E stride=16 -b 4096' --mgsnode=m...@tcp0 /dev/sdb now upgrading: tunefs.lustre --param=failover.mode=failout --fsname lustre --ost --mgsnode=m...@tcp0 /dev/sdb You actually don't need to run tunefs unless you are changing something. Additionally, all parameters should be sticky across tunefs; you can see the current settings with 'tunefs --print', add new parameters with 'tunefs --param', and erase all old params with 'tunefs --erase'. A -writeconf is not intended during the upgrade. In that case, most tunefs changes won't impact anything anyhow :) How long will it take (approx) to tunefs an 8TB OST partition ? No time. There's nothing that is done to the disk format during tunefs, just the modification of a single config file. Thanks and Regards Heiko ___ Lustre-discuss mailing list Lustre-discuss@lists.lustre.org http://lists.lustre.org/mailman/listinfo/lustre-discuss ___ Lustre-discuss mailing list Lustre-discuss@lists.lustre.org http://lists.lustre.org/mailman/listinfo/lustre-discuss
Re: [Lustre-discuss] Help with a question about lustre_rsync in Lustre 2.0 Beta1
No. This tool is intended to keep an entire (big) FS synched; if all you need is a subdir you should use regular (not lustre-) rsync. (I can't think of any reason why that functionality couldn't be added, but it's not entirely trivial.) On May 19, 2010, at 8:32 AM, Yujun Wu wrote: Hello, Could somebody help me with a question with the new lustre_rsync feature? I saw from the example that lustre_rsync can synchronize a Lustre file system (/mnt/lustre) to a target file system (/mnt/target): $ lustre_rsync --source=/mnt/lustre --target=/mnt/target .. , Can I just rsync a subdirectory like: $ lustre_rsync --source=/mnt/lustre/important_data --target=/mnt/target/backup ... ? Thanks in advance for answering my question. Yujun ___ Lustre-discuss mailing list Lustre-discuss@lists.lustre.org http://lists.lustre.org/mailman/listinfo/lustre-discuss ___ Lustre-discuss mailing list Lustre-discuss@lists.lustre.org http://lists.lustre.org/mailman/listinfo/lustre-discuss
Re: [Lustre-discuss] 1.6.7.2 Replacing an OST
You can, but in general this is probably not a good idea. The filesystem will be looking for stripes for the old ost here, and they won't be there. If there were no stripes on the old ost, you can mkfs --index and writeconf everywhere. Writeconf will clear the conf_param deactivation. On Mar 30, 2010, at 12:37 PM, Scott Barber sc...@imemories.com wrote: All servers are 1.6.7.2 CentOS 5 x86_64 We had an OST that we removed previously and had been marked as inactive by: lctl --device number deactivate and then lctl conf_param name.osc.active=0 If I have a new OST to add to the system can I format it to replace the removed OST? Can I pass mkfs.lustre the old OST name (ie: OST000c)? If so do I have to reactivate the OST? Thanks, Scott Barber Senior System Administrator iMemories.com ___ Lustre-discuss mailing list Lustre-discuss@lists.lustre.org http://lists.lustre.org/mailman/listinfo/lustre-discuss ___ Lustre-discuss mailing list Lustre-discuss@lists.lustre.org http://lists.lustre.org/mailman/listinfo/lustre-discuss