HTCondor Annex User’s Guide¶
A user of condor_annex may be a regular job submitter, or she may be an HTCondor pool administrator. This guide will cover basic condor_annex usage first, followed by advanced usage that may be of less interest to the submitter. Users interested in customizing condor_annex should consult the HTCondor Annex Customization Guide.
Considerations and Limitations¶
When you run condor_annex, you are adding (virtual) machines to an HTCondor pool. As a submitter, you probably don’t have permission to add machines to the HTCondor pool you’re already using; generally speaking, security concerns will forbid this. If you’re a pool administrator, you can of course add machines to your pool as you see fit. By default, however, condor_annex instances will only start jobs submitted by the user who started the annex, so pool administrators using condor_annex on their users’ behalf will probably want to use the -owners option or -no-owner flag; see the condor_annex man page. Once the new machines join the pool, they will run jobs as normal.
Submitters, however, will have to set up their own personal HTCondor pool, so that condor_annex has a pool to join, and then work with their pool administrator if they want to move their existing jobs to their new pool. Otherwise, jobs will have to be manually divided (removed from one and resubmitted to the other) between the pools. For instructions on creating a personal HTCondor pool, preparing an AWS account for use by condor_annex, and then configuring condor_annex to use that account, see the Using condor_annex for the First Time section.
Starting in v8.7.1, condor_annex will check for inbound access to the collector (usually port 9618) before starting an annex (it does not support other network topologies). When checking connectivity from AWS, the IP(s) used by the AWS Lambda function implementing this check may not be in the same range(s) as those used by AWS instance; please consult AWS’s list of all their IP 2 when configuring your firewall.
Starting in v8.7.2, condor_annex requires that the AWS secret (private) key file be owned by the submitting user and not readable by anyone else. This helps to ensure proper attribution.
This section assumes you’re logged into a Linux machine an that you’ve already configured condor_annex. If you haven’t, see the Using condor_annex for the First Time section.
All the terminal commands (shown in a box without a title) and file edits (shown in a box with an emphasized filename for a title) in this section take place on the Linux machine. In this section, we follow the common convention that the commands you type are preceded by by ‘$’ to distinguish them from any expected output; don’t copy that part of each of the following lines. (Lines which end in a ‘\’ continue on the following line; be sure to copy both lines. Don’t copy the ‘\’ itself.)
What You’ll Need to Know¶
To create a HTCondor annex with on-demand instances, you’ll need to know two things:
A name for it. “MyFirstAnnex” is a fine name for your first annex.
How many instances you want. For your first annex, when you’re checking to make sure things work, you may only want one instance.
Start an Annex¶
Entering the following command will start an annex named “MyFirstAnnex” with one instance. condor_annex will print out what it’s going to do, and then ask you if that’s OK. You must type ‘yes’ (and hit enter) at the prompt to start an annex; if you do not, condor_annex will print out instructions about how to change whatever you may not like about what it said it was going to do, and then exit.
$ condor_annex -count 1 -annex-name MyFirstAnnex Will request 1 m4.large on-demand instance for 0.83 hours. Each instance will terminate after being idle for 0.25 hours. Is that OK? (Type 'yes' or 'no'): yes Starting annex... Annex started. Its identity with the cloud provider is 'TestAnnex0_f2923fd1-3cad-47f3-8e19-fff9988ddacf'. It will take about three minutes for the new machines to join the pool.
You won’t need to know the annex’s identity with the cloud provider unless something goes wrong.
Before starting the annex, condor_annex (v8.7.1 and later) will check to make sure that the instances will be able to contact your pool. Contact the Linux machine’s administrator if condor_annex reports a problem with this step.
By default, condor_annex arranges for your annex’s instances to be terminated after 0.83 hours (50 minutes) have passed. Once it’s in place, this lease doesn’t depend on the Linux machine, but it’s only checked every five minutes, so give your deadlines a lot of cushion to make you don’t get charged for an extra hour. The lease is intended to help you conserve money by preventing the annex instances from accidentally running forever. You can specify a lease duration (in decimal hours) with the -duration flag.
If you need to adjust the lease for a particular annex, you may do so by specifying an annex name and a duration, but not a count. When you do so, the new duration is set starting at the current time. For example, if you’d like “MyFirstAnnex” to expire eight hours from now:
$ condor_annex -annex-name MyFirstAnnex -duration 8 Lease updated.
By default, condor_annex will configure your annex’s instances to terminate themselves after being idle for 0.25 hours (fifteen minutes). This is intended to help you conserve money in case of problems or an extended shortage of work. As noted in the example output above, you can specify a max idle time (in decimal hours) with the -idle flag. condor_annex considers an instance idle if it’s unclaimed (see condor_startd Policy Configuration for a definition), so it won’t get tricked by jobs with long quiescent periods.
Tagging your Annex’s Instances¶
By default, condor_annex adds a tag,
htcondor:AnnexName, to each
instance in the annex; its value is the annex’s name (as entered on the
command line). You may add additional tags via the command-line option
-tag, which must be followed by a tag name and a value for that tag
(as separate arguments). You may specify any number of tags (up to the
maximum supported by the cloud provider) by adding additional
options to the command line.
Starting Multiple Annexes¶
You may have up to fifty (or fewer, depending what else you’re doing with your AWS account) differently-named annexes running at the same time. Running condor_annex again with the same annex name before stopping that annex will both add instances to it and change its duration. Only instances which start up after an invocation of condor_annex will respect that invocation’s max idle time. That may include instances still starting up from your previous (first) invocation of condor_annex, so be sure your instances have all joined the pool before running condor_annex again with the same annex name if you’re changing the max idle time. Each invocation of condor_annex requests a certain number of instances of a given type; you may specify the instance type, the count, or both with each invocation, but doing so does not change the instance type or count of any previous request.
Monitor your Annex¶
You can find out if an instance has successfully joined the pool in the following way:
$ condor_annex status Name OpSys Arch State Activity Load firstname.lastname@example.org LINUX X86_64 Unclaimed Benchmarking 0.0 email@example.com LINUX X86_64 Unclaimed Idle 0.0 Total Owner Claimed Unclaimed Matched Preempting Backfill Drain X86_64/LINUX 2 0 0 2 0 0 0 0 Total 2 0 0 2 0 0 0 0
This example shows that the annex instance you requested has joined your pool. (The default annex image configures one static slot for each CPU it finds on start-up.)
You may instead use condor_status:
$ condor_status -annex MyFirstAnnex firstname.lastname@example.org LINUX X86_64 Unclaimed Idle 0.640 3767 email@example.com LINUX X86_64 Unclaimed Idle 0.640 3767 Total Owner Claimed Unclaimed Matched Preempting Backfill Drain X86_64/LINUX 2 0 0 2 0 0 0 0 Total 2 0 0 2 0 0 0 0
You can also get a report about the instances which have not joined your pool:
$ condor_annex -annex MyFirstAnnex -status STATE COUNT pending 1 TOTAL 1 Instances not in the pool, grouped by state: pending i-06928b26786dc7e6e
Monitoring Multiple Annexes¶
The following command reports on all annex instance which have joined the pool, regardless of which annex they’re from:
$ condor_status -annex firstname.lastname@example.org LINUX X86_64 Unclaimed Idle 0.640 3767 email@example.com LINUX X86_64 Unclaimed Idle 0.640 3767 firstname.lastname@example.org LINUX X86_64 Unclaimed Idle 0.640 3767 email@example.com LINUX X86_64 Unclaimed Idle 0.640 3767 Total Owner Claimed Unclaimed Matched Preempting Backfill Drain X86_64/LINUX 4 0 0 4 0 0 0 0 Total 4 0 0 4 0 0 0 0
The following command reports about instance which have not joined the pool, regardless of which annex they’re from:
$ condor_annex -status NAME TOTAL running NamelessTestA 2 2 NamelessTestB 3 3 NamelessTestC 1 1 NAME STATUS INSTANCES... NamelessTestA running i-075af9ccb40efb162 i-0bc5e90066ed62dd8 NamelessTestB running i-02e69e85197f249c2 i-0385f59f482ae6a2e i-06191feb755963edd NamelessTestC running i-09da89d40cde1f212
The ellipsis in the last column (INSTANCES…) is to indicate that it’s a very wide column and may wrap (as it has in the example), not that it has been truncated.
The following command combines these two reports:
$ condor_annex status Name OpSys Arch State Activity Load firstname.lastname@example.org LINUX X86_64 Unclaimed Benchmarking 0.0 email@example.com LINUX X86_64 Unclaimed Idle 0.0 Total Owner Claimed Unclaimed Matched Preempting Backfill Drain X86_64/LINUX 2 0 0 2 0 0 0 0 Total 2 0 0 2 0 0 0 0 Instance ID not in Annex Status Reason (if known) i-075af9ccb40efb162 NamelessTestA running - i-0bc5e90066ed62dd8 NamelessTestA running - i-02e69e85197f249c2 NamelessTestB running - i-0385f59f482ae6a2e NamelessTestB running - i-06191feb755963edd NamelessTestB running - i-09da89d40cde1f212 NamelessTestC running -
Run a Job¶
Starting in v8.7.1, the default behaviour for an annex instance is to run only jobs submitted by the user who ran the condor_annex command. If you’d like to allow other users to run jobs, list them (separated by commas; don’t forget to include yourself) as arguments to the -owner flag when you start the instance. If you’re creating an annex for general use, use the -no-owner flag to run jobs from anyone.
Also starting in v8.7.1, the default behaviour for an annex instance is
to run only jobs which have the MayUseAWS attribute set (to true). To
submit a job with MayUseAWS set to true, add
+MayUseAWS = TRUE to the
submit file somewhere before the queue command. To allow an existing job
to run in the annex, use condor_q_edit. For instance, if you’d like
cluster 1234 to run on AWS:
$ condor_qedit 1234 "MayUseAWS = TRUE" Set attribute "MayUseAWS" for 21 matching jobs.
Stop an Annex¶
The following command shuts HTCondor off on each instance in the annex; if you’re using the default annex image, doing so causes each instance to shut itself down. HTCondor does not provide a direct method terminating condor_annex instances.
$ condor_off -annex MyFirstAnnex Sent "Kill-Daemon" command for "master" to master ip-172-31-48-84.ec2.internal
Stopping Multiple Annexes¶
The following command turns off all annex instances in your pool, regardless of which annex they’re from:
$ condor_off -annex Sent "Kill-Daemon" command for "master" to master ip-172-31-48-84.ec2.internal Sent "Kill-Daemon" command for "master" to master ip-111-48-85-13.ec2.internal
Using Different or Multiple AWS Regions¶
It sometimes advantageous to use multiple AWS regions, or convenient to
use an AWS region other than the default, which is
us-east-1. To change
the default, set the configuration macro ANNEX_DEFAULT_AWS_REGION
to the new default. (If you used
the condor_annex automatic setup, you can edit the
.condor directory in your home directory; this file uses the normal
HTCondor configuration file syntax. (See
Ordered Evaluation to Set the Configuration.) Once you do this, you’ll
have to re-do the setup, as setup is region-specific.
If you’d like to use multiple AWS regions, you can specify which reason to use on the command line with the -aws-region flag. Each region may have zero or more annexes active simultaneously.
The previous section covered using what AWS calls “on-demand” instances. (An “instance” is “a single occurrence of something,” in this case, a virtual machine. The intent is to distinguish between the active process that’s pretending to be a real piece of hardware - the “instance” - and the template it used to start it up, which may also be called a virtual machine.) An on-demand instance has a price fixed by AWS; once acquired, AWS will let you keep it running as long as you continue to pay for it.
In constrast, a “Spot” instance has a price determined by an (automated) auction; when you request a “Spot” instance, you specify the most (per hour) you’re willing to pay for that instance. If you get an instance, however, you pay only what the spot price is for that instance; in effect, AWS determines the spot price by lowering it until they run out of instances to rent. AWS advertises savings of up to 90% over on-demand instances.
There are two drawbacks to this cheaper type of instance: first, you may have to wait (indefinitely) for instances to become available at your preferred price-point; the second is that your instances may be taken away from you before you’re done with them because somebody else will pay more for them. (You won’t be charged for the hour in which AWS kicks you off an instance, but you will still owe them for all of that instance’s previous hours.) Both drawbacks can be mitigated (but not eliminated) by bidding the on-demand price for an instance; of course, this also minimizes your savings.
Determining an appropriate bidding strategy is outside the purview of this manual.
Using AWS Spot Fleet¶
condor_annex supports Spot instances via an AWS technology called “Spot Fleet”. Normally, when you request instances, you request a specific type of instance (the default on-demand instance is, for instance, ‘m4.large’.) However, in many cases, you don’t care too much about how many cores an intance has - HTCondor will automatically advertise the right number and schedule jobs appropriately, so why would you? In such cases - or in other cases where your jobs will run acceptably on more than one type of instance - you can make a Spot Fleet request which says something like “give me a thousand cores as cheaply as possible”, and specify that an ‘m4.large’ instance has two cores, while ‘m4.xlarge’ has four, and so on. (The interface actually allows you to assign arbitrary values - like HTCondor slot weights - to each instance type 1, but the default value is core count.) AWS will then divide the current price for each instance type by its core count and request spot instances at the cheapest per-core rate until the number of cores (not the number of instances!) has reached a thousand, or that instance type is exhausted, at which point it will request the next-cheapest instance type.
(At present, a Spot Fleet only chooses the cheapest price within each AWS region; you would have to start a Spot Fleet in each AWS region you were willing to use to make sure you got the cheapest possible price. For fault tolerance, each AWS region is split into independent zones, but each zone has its own price. Spot Fleet takes care of that detail for you.)
In order to create an annex via a Spot Fleet, you’ll need a file containing a JSON blob which describes the Spot Fleet request you’d like to make. (It’s too complicated for a reasonable command-line interface.) The AWS web console can be used to create such a file; the button to download that file is (currently) in the upper-right corner of the last page before you submit the Spot Fleet request; it is labeled ‘JSON config’. You may need to create an IAM role the first time you make a Spot Fleet request; please do so before running condor_annex.
You must select the instance role profile used by your on-demand instances for condor_annex to work. This value will have been stored in the configuration macro
ANNEX_DEFAULT_ODI_INSTANCE_PROFILE_ARNby the setup procedure.
You must select a security group which allows inbound access on HTCondor’s port (9618) for condor_annex to work. You may use the value stored in the configuration macro
ANNEX_DEFAULT_ODI_SECURITY_GROUP_IDSby the setup procedure; this security group also allows inbound SSH access.
If you wish to be able to SSH to your instances, you must select an SSH key pair (for which you have the corresponding private key); this is not required for condor_ssh_to_job. You may use the value stored in the configuration macro
ANNEX_DEFAULT_ODI_KEY_NAMEby the setup procedure.
Specify the JSON configuration file using -aws-spot-fleet-config-file, or set the configuration macro ANNEX_DEFAULT_SFR_CONFIG_FILE to the full path of the file you just downloaded, if you’d like it to become your default configuration for Spot annexes. Be aware that condor_annex does not alter the validity period if one is set in the Spot Fleet configuration file. You should remove the references to ‘ValidFrom’ and ‘ValidTo’ in the JSON file to avoid confusing surprises later.
Additionally, be aware that condor_annex uses the Spot Fleet API in its “request” mode, which means that an annex created with Spot Fleet has the same semantics with respect to replacement as it would otherwise: if an instance terminates for any reason, including AWS taking it away to give to someone else, it is not replaced.
You must specify the number of cores (total instance weight; see above) using -slots. You may also specify -aws-spot-fleet, if you wish; doing so may make this condor_annex invocation more self-documenting. You may use other options as normal, excepting those which begin with -aws-on-demand, which indicates an option specific to on-demand instances.
Custom HTCondor Configuration¶
When you specify a custom configuration, you specify the full path to a configuration directory which will be copied to the instance. The customizations performed by condor_annex will be applied to a temporary copy of this directory before it is uploaded to the instance. Those customizations consist of creating two files: password_file.pl (named that way to ensure that it isn’t ever accidentally treated as configuration), and 00ec2-dynamic.config. The former is a password file for use by the pool password security method, which if configured, will be used by condor_annex automatically. The latter is an HTCondor configuration file; it is named so as to sort first and make it easier to over-ride with whatever configuration you see fit.
AWS Instance User Data¶
HTCondor doesn’t interfere with this in any way, so if you’d like to set an instance’s user data, you may do so. However, as of v8.7.2, the -user-data options don’t work for on-demand instances (the default type). If you’d like to specify user data for your Spot Fleet -driven annex, you may do so in four different ways: on the command-line or from a file, and for all launch specifications or for only those launch specifications which don’t already include user data. These two choices correspond to the absence or presence of a trailing -file and the absence or presence of -default immediately preceding -user-data.
A “launch specification,” in this context, means one of the virtual machine templates you told Spot Fleet would be an acceptable way to accomodate your resource request. This usually corresponds one-to-one with instance types, but this is not required.
The condor_annex manual page lists the “expert mode” options.
Four of the “expert mode” options set the URLs used to access AWS services, not including the CloudFormation URL needed by the -setup flag. You may change the CloudFormation URL by changing the HTCondor configuration macro ANNEX_DEFAULT_CF_URL , or by supplying the URL as the third parameter after the -setup flag. If you change any of the URLs, you may need to change all of the URLs - Lambda functions and CloudWatch events in one region don’t work with instances in another region.
You may also temporarily specify a different AWS account by using the access (-aws-access-key-file) and secret key (-aws-secret-key-file) options. Regular users may have an accounting reason to do this.
The options labeled “developers only” control implementation details and may change without warning; they are probably best left unused unless you’re a developer.