You are here

CFDR Data

The LANL data

System: 22 HPC cluster systems
Duration: December 1996 thru November 2005
Data Type: Records of cluster node outages, workload logs and error logs.

About the data:

The data spans 22 high-performance computing systems that have been in production use at Los Alamos National Lab (LANL) between 1996 and November 2005. Most of these systems are large clusters of either NUMA (Non-Uniform-Memory-Access) nodes, or 2-way and 4-way SMP (Symmetric-Multi-Processing) nodes. In total the systems include 4750 nodes and 24101 processors.

The data contains an entry for any failure leading to a node outage that occurred during the 9-year time period and that required the attention of a system administrator. For each failure, the data includes start time and end time, the system and node affected, as well as categorized root cause information.

The workloads run on those systems are large-scale long-running 3D scientific simulations, e.g. for nuclear stockpile stewardship. These applications perform long periods (often months) of CPU computation, interrupted every few hours by a few minutes of I/O for checkpointing. Simulation workloads are often accompanied by scientific visualization of large-scale data. Visualization workloads are also CPU-intensive, but involve more reading from storage than compute workloads. Finally, some nodes are used purely as front-end nodes, and others run more than one type of workload, e.g. graphics nodes often run compute workloads as well.

At LANL, failure tolerance is frequently implemented through periodic checkpointing. When a node fails, the job(s) running on it is stopped and restarted on a different set of nodes, either starting from the most recent checkpoint or from scratch if no checkpoint exists.

The data is based on a ``remedy'' database created at LANL in June 1996. At that time, LANL introduced a site-wide policy that requires system administors to enter a description of every failure they take care of into the remedy database. Consequentially, the database contains a record for every failure that occurred in LANL's HPC systems since June 1996 and that required intervention of a system administrator.

A failure record contains the time when the failure started, the time when it was resolved, the system and node affected, the type of workload running on the node and the root cause. The workload is either compute for computational workloads, graphics for visualization workloads, or fe for front-end. Root causes fall in one of the following five high-level categories: Human error; Environment, including power outages or A/C failures; Network failure; Software failure; and Hardware failure. In addition, more detailed information on the root cause is captured, such as the particular hardware component affected by a Hardware failure. The failure classification and rules for assigning failures to categories were developed jointly by hardware engineers, administrators and operations staff.

The LANL web page and a recent paper provide more detailed information about the systems the data was collected on and the data format and a FAQ.

Downloads:

The data can be downloaded directly from the LANL web page. The LANL web page also provides more detailed information on the data format and the systems the data was collected on. There is also a FAQ available.

Papers using this data:

Bianca Schroeder and Garth A. Gibson. "A large scale study of failures in high-performance-computing systems." International Symposium on Dependable Systems and Networks (DSN 2006).
Gonzalo Zarza, Diego Lugones, Daniel Franco and Emilio Luque. "Fault-tolerant Routing for Multiple Permanent and Non-permanent Faults in HPC Systems." 2010 International Conference on Parallel and Distributed Processing Techniques and Applications (PDPTA'10), Las Vegas, Nevada, USA, July 2010.

If you are using this data in a paper, please send an e-mail with the paper reference to the moderators and we will add it to this page.

Acknowledgments:

We thank Gary Grider, Laura Davey, James Nunez, and the Computing, Communications, and Networking Division at LANL for their efforts in collecting the data and clearing it for public release and for their help with background information and interpretation.

If you use these data in your work, please use a similar acknowledgment.


The HPC1 data

System: A 765-node HPC cluster with 64 filesystem nodes
Duration: August 2001 thru May 2006
Data Type: Hardware replacement log

About the data:

HPC1 is a five year log of hardware replacements collected from a 765 node high-performance computing cluster. Each of the 765 nodes is a 4-way SMP with 4 GB of memory and three to four 18GB 10K rpm SCSI drives. Of these nodes, 64 are used as filesystem nodes containing, in addition to the three to four 18GB drives, 17 36GB 10K rpm SCSI drives. The applications running on this system are typically large-scale scientific simulations or visualization applications. The data contains, for each hardware replacement that was recorded during the five year lifetime of this system, when the problem started, which node and which hardware component was affected, and a brief description of the corrective action.

Downloads:

After registering you may download the data set here (test only; not real data).

Papers using this data:

A first analysis of the HPC1 data is presented in the following paper:
Bianca Schroeder and Garth A. Gibson. "Disk failures in the real world: What does an MTTF of 1,000,000 hours mean too you?". 5th USENIX Conference on File and Storage Technologies (FAST 2007).

If you are using this data in a paper, please send an e-mail with the paper reference to the moderators and we will add it to this page.

Acknowledgments:

We would like to thank Katie Vargo, J. Ray Scott and Robin Flaus from the Pittsburgh Supercomputing Center for collecting and providing us with data and helping us to interpret the data.

If you use these data in your work, please use a similar acknowledgment.


The HPC2 data

System: A 256-node HPC cluster
Duration: January 2004 thru July 2006
Data Type: Hardware replacement log

About the data:

HPC2 is a record of disk replacements observed on the compute nodes of a 256 node HPC cluster at Los Alamos National Lab (LANL). Each node is a 4-way SMP with 16 GB of memory and contains two 36GB 10K rpm SCSI drives, except for eight of the nodes, which contain eight 36GB 10K rpm SCSI drives each. The applications running on this system are typically large-scale scientific simulations or visualization applications. For each disk replacement, the data set records the number of the affected node, the start time of the problem, and the slot number of the replaced drive.

Downloads:

The data can be downloaded directly from the LANL web page. The LANL web page also provides more detailed information on the data format and the system the data was collected on.

Papers using this data:

A first analysis of the HPC2 data is presented in the following paper:
Bianca Schroeder and Garth A. Gibson. "Disk failures in the real world: What does an MTTF of 1,000,000 hours mean too you?". 5th USENIX Conference on File and Storage Technologies (FAST 2007).

If you are using this data in a paper, please send an e-mail with the paper reference to the moderators and we will add it to this page.

Acknowledgments:

We thank Gary Grider, Laura Davey, James Nunez, and the Computing, Communications, and Networking Division at LANL for their efforts in collecting the data and clearing it for public release and for their help with background information and interpretation.

If you use these data in your work, please use a similar acknowledgment.


The HPC3 data

System: 1,532 node HPC cluster
Duration: December 2005 thru November 2006
Data Type: Harddrive replacement log

About the data:

HPC3 is a record of disk replacements observed on a 1,532 node HPC cluster. Each node is equipped with eight CPUs and 32GB of memory. Each node, except for four login nodes, has two 146GB 15K rpm SCSI disks. In addition, 11,000 7200 rpm 250GB SATA drives are used in an external shared filesystem and 144 73GB 15K rpm SCSI drives are used for the filesystem metadata. The applications running on this system are typically large-scale scientific simulations or visualization applications. For each disk replacement, the data set records the day of the replacement.

Downloads:

The data will soon become available for download.

Papers using this data:

A first analysis of the HPC3 data is presented in the following paper:
Bianca Schroeder and Garth A. Gibson. "Disk failures in the real world: What does an MTTF of 1,000,000 hours mean too you?". 5th USENIX Conference on File and Storage Technologies (FAST 2007).

If you are using this data in a paper, please send an e-mail with the paper reference to the moderators and we will add it to this page.

Acknowledgments:

We would like to thank the people at the organization, who has provided us with data, but would like to remain unnamed, for collecting the data and helping us to interpret the data.


The HPC4 data

System: Five HPC systems with 512 to 131072 processors
Duration: Between 215 and 558 days in 2004 - 2006
Data Type: Event logs

About the data:

For a detailed description of the data see the DSN'07 paper by Adam Oliner and Jon Stearley. Below follows a very brief summary. The data consists of five event logs collected between 2004 and 2006 on five different supercomputing systems: Blue Gene/L, Thunderbird, RedStorm, Spirit, and Liberty. The logs contain alert and non-alert messages identified by alert category tags, and are therefore more amenable to alert detection and prediction research than failure modeling. All five systems are ranked on the Top500 Supercomputers List as of June 2006, spanning a range from #1 to #445. They vary by two orders of magnitude in the number of processors (ranging from 512 proccesors in Liberty to 131072 processors in Blue Gene/L) and by one order of magnitude in the amount of main memory The various machines are produced by IBM, Dell, Cray and HP. All systems are installed at Sandia National Labs (SNL) in Albuquerque, NM, with the exception of BG/L, which is at Lawrence Livermore National Labs (LLNL) in Livermore, CA.

Downloads:

Thunderbird (1.9 GB) | Spirit (864 MB) | Liberty (641 MB) | BlueGene/L (60 MB)

Papers using this data:

A description and analysis of the HPC4 data is presented in the following paper:

Oliner and J. Stearley. "What Supercomputers Say: A Study of Five System Logs". Proceedings of the International Conference on Dependable Systems and Networks (DSN), Edinburgh, UK, 2007.

The data is also used in the following two papers:

A. J. Oliner, A. Aiken, and J. Stearley. "Alert Detection in System Logs" Proceedings of the International Conference on Data Mining (ICDM), Pisa, Italy, 2008.

J. Stearley and A. J. Oliner. "Bad Words: Finding Faults in Spirit’s Syslogs.". Workshop on Resiliency in High-Performance Computing (Resilience-2008), Lyon, France, 2008.

If you are using this data in a paper, please send an e-mail with the paper reference to the moderators and we will add it to this page.

Acknowledgments:

We would like to thank Jon Stearley and Adam Oliner for collecting the data and making it publicly available.


The PNNL data

System: 980 node HPC cluster
Duration: November 2003 thru September 2007
Data Type: Log of hardware failures

About the data:

This data set is a record of hardware failures recorded on the High Performance Computing System-2 (MPP2) operated by the Environmental and Molecular Science Labratory (EMSL), Molecular Science Computing Facility (MSCF) at Pacific Northwest National Laboratory (PNNL).

The MPP2 computing system has the following equipment and capabilities:

  • HP/Linux Itanium-2
  • 980 node/1960 Itanium-2 processors (Madison, 1.5 GHz) configured as follows:
    • 574 nodes are "fat" compute nodes with 10 Gbyte RAM and 430 Gbyte local disk
    • 366 nodes are "thin" compute nodes with 10 Gbyte RAM and 10 Gbyte local disk
    • 34 nodes are Lustre server nodes (32 OSS, 2 MDS)
    • 2 nodes are administrative nodes
    • 4 nodes are login nodes
  • Quadrics QsNetII interconnect
  • 11.8 TFlops peak theoretical performance
  • 9.7 terabytes of RAM
  • 450 terabytes of local scratch disk space
  • 53 terabytes shared cluster file system (Lustre).

     

The applications running on this system are typically large-scale scientific simulations or visualization applications.

For each hardware failure, the data set includes a timestamp for when the failure happened, a hardware identifier, the component that failed, a description of the failure, and the repair action taken.

Downloads:

The data is available for download here.

Papers using this data:

Gonzalo Zarza, Diego Lugones, Daniel Franco and Emilio Luque. "Fault-tolerant Routing for Multiple Permanent and Non-permanent Faults in HPC Systems." 2010 International Conference on Parallel and Distributed Processing Techniques and Applications (PDPTA'10), Las Vegas, Nevada, USA, July 2010.

If you are using this data in a paper, please send an e-mail with the paper reference to the moderators and we will add it to this page.

Acknowledgments:

We would like to thank Evan Felix and David Brown from PNNL for collecting the data and sharing it. The data was collected and made available using the Molecular Science Computing Facility (MSCF) in the William R. Wiley Environmental Molecular Sciences Laboratory, a national scientific user facility sponsored by the U.S. Department of Energy's Office of Biological and Environmental Research and located at the Pacific Northwest National Laboratory, operated for the Department of Energy by Battelle.

If you use these data in your work, please use a similar aedgment.


The NERSC data

System: Various HPC clusters
Duration: 2001 - 2006
Data Type: Database with I/O specific failures

About the data:

This data was collected with the purpose of providing failure specifics for I/O related systems and components in as much detail as possible so that analysis might produce some useful findings. Data were collected for storage, networking, computational machines, and file systems in production use at NERSC from the 2001-2006 timeframe. The data was extracted form a database used for tracking system troubles, called Remedy, and is currently stored in a mySQL database and available for export to Excel format. There are also some basic query and graph capabilities available. For more information on the data, please visit the NERSC web site hosting the raw data or contact the PDSI researcher at NERSC: Akbar Mokhtarani or the Principal Investigator for PDSI at NERSC: Bill Kramer .

Downloads:

The data and more information is available for download here.

Papers using this data:

This data has not yet been reported on in any paper.

If you are using this data in a paper, please send an e-mail with the paper reference to the moderators and we will add it to this page.

Acknowledgments:

We would like to thank Bill Kramer and Akbar Mokhtarani from NERSC for collecting the data and sharing it.

If you use these data in your work, please use a similar acknowledgment.


The COM1 data

System: Internet services clusters
Duration: May 2006
Data Type: Hardware replacement log

About the data:

COM1 is a log of hardware failures recorded by an internet service provider and drawing from multiple distributed sites. Each record in the data contains a timestamp of when the failure was repaired, information on the failure symptoms, and a list of steps that were taken to diagnose and repair the problem. The data does not contain information on when each failure actually happened, only when repair took place. The data covers a population of 26,734 10K rpm SCSI disk drives. The total number of servers in the monitored sites is not known.

Downloads:

The data will soon become available for download.

Papers using this data:

A first analysis of the HPC3 data is presented in the following paper:
Bianca Schroeder and Garth A. Gibson. "Disk failures in the real world: What does an MTTF of 1,000,000 hours mean too you?". 5th USENIX Conference on File and Storage Technologies (FAST 2007).

If you are using this data in a paper, please send an e-mail with the paper reference to the moderators and we will add it to this page.

Acknowledgments:

We would like to thank the people at the organization, who has provided us with data, but would like to remain unnamed, for collecting the data and helping us to interpret the data.


The COM2 data

System: Internet services cluster
Duration: September 2004 thru April 2006
Data Type: Warranty service log of hardware failures

About the data:

COM2 is a warranty service log of hardware failures recorded on behalf of an internet service provider aggregating events in multiple distributed sites. Each failure record contains a repair code (e.g. ``Replace hard drive'') and the time when the repair was finished. Again there is no information on the start time of each failure. The log does not contain entries for failures of disks that were replaced in the customer site by hot-swapping in a spare disk, since the data was created by the warranty processing, which does not participate in on-site hot-swap replacements. To account for the missing disk replacements we obtained numbers for the periodic replenishments of on-site spare disks from the internet service provider. The size of the underlying system changed significantly during the measurement period, starting with 420 servers in 2004 and ending with 9,232 servers in 2006. We obtained quarterly hardware purchase records covering this time period that make it possible to estimate the size of the disk population.

Downloads:

The data will soon become available for download.

Papers using this data:

A first analysis of the HPC3 data is presented in the following paper:
Bianca Schroeder and Garth A. Gibson. "Disk failures in the real world: What does an MTTF of 1,000,000 hours mean too you?". 5th USENIX Conference on File and Storage Technologies (FAST 2007).

If you are using this data in a paper, please send an e-mail with the paper reference to the moderators and we will add it to this page.

Acknowledgments:

We would like to thank the people at the organization, who has provided us with data, but would like to remain unnamed, for collecting the data and helping us to interpret the data.


The COM3 data

System: Internet services clusters
Duration: January 2005 thru December 2005
Data Type: Aggregate harddrive replacement statistics

About the data:

The COM3 data set comes from a large external storage system used by an internet service provider and comprises four populations of different types of FC disks. While this data was gathered in 2005, the system has some legacy components that were as old as from 1998 and were known to have been physically moved after initial installation. COM3 differs from the other data sets in that it provides only aggregate statistics of disk failures, rather than individual records for each failure. The data contains the counts of disks that failed and were replaced in 2005 for each of the four disk populations.

Downloads:

The data will soon become available for download.

Papers using this data:

A first analysis of the HPC3 data is presented in the following paper:
Bianca Schroeder and Garth A. Gibson. "Disk failures in the real world: What does an MTTF of 1,000,000 hours mean too you?". 5th USENIX Conference on File and Storage Technologies (FAST 2007).

If you are using this data in a paper, please send an e-mail with the paper reference to the moderators and we will add it to this page.

Acknowledgments:

We would like to thank the people at the organization, who has provided us with data, but would like to remain unnamed, for collecting the data and helping us to interpret the data.


 

The Cray data

System: Cray system
Duration: N/A
Data Type: Event logs, syslog, console logs

About the data:

These data sets come from one or more Cray XT series machines, running Linux. They include the syslog, the event log and the console log. The syslog contains messages produced by various Linux daemons, drivers and utilities that use the syslog protocol. The event log is an XT-specific log that records actions on the XT control network, such as transfers of boot images. The console log is the aggregate console output of all the nodes; the console log may partially overlap with the syslog and event log.

In addition to the system events, we will also provide a failure log that will contain a description of the failures that were encountered. This will include information about which log entry was critical in identifying this error if this information was known.

The files for download are .tar.gz files, that when unpacked, will have a README file in it. The first section of the README explains what Cray thought happened and (very) briefly why it was thought that was the problem. There are 6 dumps. The directory name is the YYMMDDHHMM that the dump started at (basically, a unique ID). The next sections are the same in all the README's. They explain a bit about how the machine is put together, the naming convention for nodes, and finally, what all the log files present in each dump are.

Downloads:

Data set 1
Data set 2
Data set 3
Data set 4
Data set 5
Data set 6

Papers using this data:

If you are using this data in a paper, please send an e-mail with the paper reference to the moderators and we will add it to this page.

Acknowledgments:

We thank Forest Godfrey at Cray for making this data available.


Blue Gene/P data from Intrepid

System: Blue Gene/P
Duration: Jan 09 - Aug 09
Data Type: RAS log

About the data:

The data consists of RAS log messages collected over a period of 6 months on the Blue Gene/P Intrepid system at. Each message in the log contains 15 fields as follows: RECID, MSG_ID, COMPONENT, SUBCOMPONENT, ERRCODE, SEVERITY, EVENT_TIME, FLAGS, PROCESSOR, NODE, BLOCK, LOCATION, SERIALNUMBER, ECID, MESSAGE.

More details about the data and system it comes from will be made available soon. In the meantime, please refer to the paper by Zheng et al. listed below which provides an analysis of the data.

Downloads:

RAS log

Papers using this data:

Z. Zheng, L. Yu, W. Tang, Z. Lan, R. Gupta, N. Desai, S. Coghlan, and D. Buettner, ''Co-Analysis of RAS Log and Job Log on Blue Gene/P,'' in Proc. of IEEE International Parallel & Distributed Processing Symposium (IPDPS'11), Anchorage, AK, USA, 2011.

Acknowledgments:

We thank Ziming Zheng and Zhiling Lan from the Illinois Institute of Technology for making this data set available.

This content is available to: