Re: [DISCUSS] About the usage of CUDA/CUDNN
Hi, sorry for derailing the discussion another time (I'm slow to respond, sorry) but I just wanted to give a little statement about the security aspect of Jenkins. Kellen is right that Jenkins is by its nature quite insecure due to the high frequency of discovered security vulnerabilities. These attacks go either against slaves due to arbitrary code execution (which we have mitigated with the restricted slaves) or against the master with its big attack surface due to high number of used plugins in combination with infrequence updates from our side. This decision to use the CI system for CD purposes was done as a result of a lacking alternative - which lead me to quickly coming up with the restricted slave idea. This was, from the beginning, intended as a temporary solution and I still stand by that. That's why Chance Bair and I are currently working on refactoring our CI deployment pipeline. This will give us three things: 1. Frequent security updates 2. Fully automated redeployment 3. Individual deployments #3 is the point which is of interest for this discussion. In future, it will be possible to deploy different environments - this could be Alpha, Beta, Gamma, Prod or - from another perspective - CI and CD. Each deployment could then also have its own security setup. What I've been thinking of (and please note that this is just some loud thinking from my side and no agreed-on or reviewed proposal) is that we could then have a CD instance of Jenkins that is not accessible from the internet or with a Nginx proxy that allows authentication before hitting the actual Jenkins Master. Before we can get to that point though, we first have to finish that CI deployment pipeline and then see where we stand. We are then going to follow up with a proposal where the community can express their ideas about the different approaches. But one thing is clear: We won't be running a mixed CI and CD forever. But Jenkins - if designed and managed properly - should not be an attack vector and thus not discarded as option for further considerations. Again, for sorry for derailing the topic afterwards. Best regards, Marco On Tue, Dec 18, 2018 at 1:07 AM Hen wrote: > Please raise on legal-discuss or in a legal jira. > > On Mon, Dec 17, 2018 at 3:59 PM Naveen Swamy wrote: > > > Attempting to answer Qing's question > > -- > > If you can digest the legal terms: > > https://docs.nvidia.com/cuda/eula/index.html#distribution-requirements. > > It sounds its OK(" > > > >1. Your application must have material additional functionality, > >beyond the included portions of the SDK.") > > > > but I not don't understand the legal lingo > > > > @Hen : Could you provide input to this? > > > > Thanks, Naveen > > > > On Mon, Dec 17, 2018 at 3:29 PM Davydenko, Denis < > > dzianis.davydze...@gmail.com> wrote: > > > >> Kellen, please see conversation [1] on previously published proposal re: > >> maven publishing pipeline. I think your concerns are valid and we should > >> look into security aspect of running our CI on a broader scope, not > bound > >> to just artifact publishing. > >> > >> I believe right now Qing's question is whether it is OK from legal > >> perspective to download CUDA by literally running wget during one of the > >> jobs in publishing pipeline. The fact it is not available by just simple > >> URL download raises concern: whether it is a protective measure from > >> downloads by unauthenticated users or just inconvenience that has not > been > >> addressed by nVidia yet. > >> > >> [1]: > >> > https://lists.apache.org/thread.html/464712f0136fb51916ca9f1b702b99847e108dbdbd0b6a2b73fc91f1@%3Cdev.mxnet.apache.org%3E > >> > >> > >> On 12/17/18, 2:48 PM, "kellen sunderland" > > >> wrote: > >> > >> Restricted nodes may provide enough security for some use cases, but > >> in my > >> opinion they don't provide enough for artifact publishing. An > example > >> would > >> be if there were a exploit available that worked against a Jenkins > >> master. > >> In this case I think an attacker code still pivot to a secure node > >> (correct > >> me if I'm wrong). > >> > >> To your second point, it shouldn't be too hard for us to maintain > all > >> the > >> deps for our packages in Dockerfiles which are checked into source > and > >> built on a regular basis. To publish these artifacts I'd recommend > >> doing > >> this from a separate, secure environment. The flow I'd recommend > >> would be > >> something like: (1) Developers commit PRs with verification that the > >> artifacts build properly on a continual basis from the CI. (2) In a > >> separate, secure environment we do the same artifact build > generation > >> again, but this time we publish to various repos as a convenience to > >> our > >> MXNet users. > >> > >> On Mon, Dec 17, 2018 at 2:34 PM Qing Lan > wrote: > >> > >> > Hi Kellen, > >> > > >> > Firstly the restricted node is completely
Re: [DISCUSS] About the usage of CUDA/CUDNN
Attempting to answer Qing's question -- If you can digest the legal terms: https://docs.nvidia.com/cuda/eula/index.html#distribution-requirements. It sounds its OK(" 1. Your application must have material additional functionality, beyond the included portions of the SDK.") but I not don't understand the legal lingo @Hen : Could you provide input to this? Thanks, Naveen On Mon, Dec 17, 2018 at 3:29 PM Davydenko, Denis < dzianis.davydze...@gmail.com> wrote: > Kellen, please see conversation [1] on previously published proposal re: > maven publishing pipeline. I think your concerns are valid and we should > look into security aspect of running our CI on a broader scope, not bound > to just artifact publishing. > > I believe right now Qing's question is whether it is OK from legal > perspective to download CUDA by literally running wget during one of the > jobs in publishing pipeline. The fact it is not available by just simple > URL download raises concern: whether it is a protective measure from > downloads by unauthenticated users or just inconvenience that has not been > addressed by nVidia yet. > > [1]: > https://lists.apache.org/thread.html/464712f0136fb51916ca9f1b702b99847e108dbdbd0b6a2b73fc91f1@%3Cdev.mxnet.apache.org%3E > > > On 12/17/18, 2:48 PM, "kellen sunderland" > wrote: > > Restricted nodes may provide enough security for some use cases, but > in my > opinion they don't provide enough for artifact publishing. An example > would > be if there were a exploit available that worked against a Jenkins > master. > In this case I think an attacker code still pivot to a secure node > (correct > me if I'm wrong). > > To your second point, it shouldn't be too hard for us to maintain all > the > deps for our packages in Dockerfiles which are checked into source and > built on a regular basis. To publish these artifacts I'd recommend > doing > this from a separate, secure environment. The flow I'd recommend > would be > something like: (1) Developers commit PRs with verification that the > artifacts build properly on a continual basis from the CI. (2) In a > separate, secure environment we do the same artifact build generation > again, but this time we publish to various repos as a convenience to > our > MXNet users. > > On Mon, Dec 17, 2018 at 2:34 PM Qing Lan wrote: > > > Hi Kellen, > > > > Firstly the restricted node is completely isolated to the > PR-checking CI > > system (physically) which is explained in here: > > > https://cwiki.apache.org/confluence/display/MXNET/Restricted+jobs+and+nodes > > . > > What you are mentioning: the Public CIs are all having troubles if > they > > are public accessible. I am not sure how secure the restricted node > is. > > However, the only way I can think of from your end is to downloading > all > > deps in a single machine and run everything there (disconnected from > > internet). It would bring us the best security we have. > > > > Thanks, > > Qing > > > > On 12/17/18, 2:06 PM, "kellen sunderland" < > kellen.sunderl...@gmail.com> > > wrote: > > > > I'm not in favour of publishing artifacts from any Jenkins based > > systems. > > There are many ways to bundle artifacts and publish them from an > > automated > > system. Why we would use a CI system like Jenkins for this task? > > Jenkins > > frequently has security vulnerabilities and is designed to run > > arbitrary > > code from the internet. It is a real possibility that an > attacker > > could > > pivot from any Jenkins based CI system to infect artifacts which > would > > then > > potentially be pushed to repositories our users would consume. > I would > > consider any system using Jenkins as insecure-by-design, and > encourage > > us > > to air-gapped any artifact generation (websites, jars, PyPi > packages) > > completely from a system like that. > > > > An alternative I could see is a simple Dockerfile (no Jenkins) > that > > builds > > all artifacts end-to-end and can be run in an automated account > well > > outside our CI account. > > > > On Mon, Dec 17, 2018 at 1:53 PM Qing Lan > wrote: > > > > > Dear community, > > > > > > Currently me and Zach are working on the Automated-publish > pipeline > > on > > > Jenkins which is a pipeline used to publish Maven packages and > pip > > packages > > > nightly build. We are trying to use NVIDIA deb which could > help us > > to build > > > different CUDA/CUDNN versions in the publish system. Sheng has > > provided a > > > script here: > https://github.com/apache/incubator-mxnet/pull/13646. > > This > > > provide a very concrete and automatic solution from > downloading to > >
Re: [DISCUSS] About the usage of CUDA/CUDNN
Kellen, please see conversation [1] on previously published proposal re: maven publishing pipeline. I think your concerns are valid and we should look into security aspect of running our CI on a broader scope, not bound to just artifact publishing. I believe right now Qing's question is whether it is OK from legal perspective to download CUDA by literally running wget during one of the jobs in publishing pipeline. The fact it is not available by just simple URL download raises concern: whether it is a protective measure from downloads by unauthenticated users or just inconvenience that has not been addressed by nVidia yet. [1]: https://lists.apache.org/thread.html/464712f0136fb51916ca9f1b702b99847e108dbdbd0b6a2b73fc91f1@%3Cdev.mxnet.apache.org%3E On 12/17/18, 2:48 PM, "kellen sunderland" wrote: Restricted nodes may provide enough security for some use cases, but in my opinion they don't provide enough for artifact publishing. An example would be if there were a exploit available that worked against a Jenkins master. In this case I think an attacker code still pivot to a secure node (correct me if I'm wrong). To your second point, it shouldn't be too hard for us to maintain all the deps for our packages in Dockerfiles which are checked into source and built on a regular basis. To publish these artifacts I'd recommend doing this from a separate, secure environment. The flow I'd recommend would be something like: (1) Developers commit PRs with verification that the artifacts build properly on a continual basis from the CI. (2) In a separate, secure environment we do the same artifact build generation again, but this time we publish to various repos as a convenience to our MXNet users. On Mon, Dec 17, 2018 at 2:34 PM Qing Lan wrote: > Hi Kellen, > > Firstly the restricted node is completely isolated to the PR-checking CI > system (physically) which is explained in here: > https://cwiki.apache.org/confluence/display/MXNET/Restricted+jobs+and+nodes > . > What you are mentioning: the Public CIs are all having troubles if they > are public accessible. I am not sure how secure the restricted node is. > However, the only way I can think of from your end is to downloading all > deps in a single machine and run everything there (disconnected from > internet). It would bring us the best security we have. > > Thanks, > Qing > > On 12/17/18, 2:06 PM, "kellen sunderland" > wrote: > > I'm not in favour of publishing artifacts from any Jenkins based > systems. > There are many ways to bundle artifacts and publish them from an > automated > system. Why we would use a CI system like Jenkins for this task? > Jenkins > frequently has security vulnerabilities and is designed to run > arbitrary > code from the internet. It is a real possibility that an attacker > could > pivot from any Jenkins based CI system to infect artifacts which would > then > potentially be pushed to repositories our users would consume. I would > consider any system using Jenkins as insecure-by-design, and encourage > us > to air-gapped any artifact generation (websites, jars, PyPi packages) > completely from a system like that. > > An alternative I could see is a simple Dockerfile (no Jenkins) that > builds > all artifacts end-to-end and can be run in an automated account well > outside our CI account. > > On Mon, Dec 17, 2018 at 1:53 PM Qing Lan wrote: > > > Dear community, > > > > Currently me and Zach are working on the Automated-publish pipeline > on > > Jenkins which is a pipeline used to publish Maven packages and pip > packages > > nightly build. We are trying to use NVIDIA deb which could help us > to build > > different CUDA/CUDNN versions in the publish system. Sheng has > provided a > > script here: https://github.com/apache/incubator-mxnet/pull/13646. > This > > provide a very concrete and automatic solution from downloading to > > installing on the system. The only scenario we are facing is: It > seemed > > NVIDIA has a restriction on distributing CUDA. We are not sure if it > is > > legally-safe for us to use this in public. > > > > We would be grateful if somebody has a better context on it and help > us > > out! > > > > Thanks, > > Qing > > > > >
Re: [DISCUSS] About the usage of CUDA/CUDNN
Restricted nodes may provide enough security for some use cases, but in my opinion they don't provide enough for artifact publishing. An example would be if there were a exploit available that worked against a Jenkins master. In this case I think an attacker code still pivot to a secure node (correct me if I'm wrong). To your second point, it shouldn't be too hard for us to maintain all the deps for our packages in Dockerfiles which are checked into source and built on a regular basis. To publish these artifacts I'd recommend doing this from a separate, secure environment. The flow I'd recommend would be something like: (1) Developers commit PRs with verification that the artifacts build properly on a continual basis from the CI. (2) In a separate, secure environment we do the same artifact build generation again, but this time we publish to various repos as a convenience to our MXNet users. On Mon, Dec 17, 2018 at 2:34 PM Qing Lan wrote: > Hi Kellen, > > Firstly the restricted node is completely isolated to the PR-checking CI > system (physically) which is explained in here: > https://cwiki.apache.org/confluence/display/MXNET/Restricted+jobs+and+nodes > . > What you are mentioning: the Public CIs are all having troubles if they > are public accessible. I am not sure how secure the restricted node is. > However, the only way I can think of from your end is to downloading all > deps in a single machine and run everything there (disconnected from > internet). It would bring us the best security we have. > > Thanks, > Qing > > On 12/17/18, 2:06 PM, "kellen sunderland" > wrote: > > I'm not in favour of publishing artifacts from any Jenkins based > systems. > There are many ways to bundle artifacts and publish them from an > automated > system. Why we would use a CI system like Jenkins for this task? > Jenkins > frequently has security vulnerabilities and is designed to run > arbitrary > code from the internet. It is a real possibility that an attacker > could > pivot from any Jenkins based CI system to infect artifacts which would > then > potentially be pushed to repositories our users would consume. I would > consider any system using Jenkins as insecure-by-design, and encourage > us > to air-gapped any artifact generation (websites, jars, PyPi packages) > completely from a system like that. > > An alternative I could see is a simple Dockerfile (no Jenkins) that > builds > all artifacts end-to-end and can be run in an automated account well > outside our CI account. > > On Mon, Dec 17, 2018 at 1:53 PM Qing Lan wrote: > > > Dear community, > > > > Currently me and Zach are working on the Automated-publish pipeline > on > > Jenkins which is a pipeline used to publish Maven packages and pip > packages > > nightly build. We are trying to use NVIDIA deb which could help us > to build > > different CUDA/CUDNN versions in the publish system. Sheng has > provided a > > script here: https://github.com/apache/incubator-mxnet/pull/13646. > This > > provide a very concrete and automatic solution from downloading to > > installing on the system. The only scenario we are facing is: It > seemed > > NVIDIA has a restriction on distributing CUDA. We are not sure if it > is > > legally-safe for us to use this in public. > > > > We would be grateful if somebody has a better context on it and help > us > > out! > > > > Thanks, > > Qing > > > > >
Re: [DISCUSS] About the usage of CUDA/CUDNN
Hi Kellen, Firstly the restricted node is completely isolated to the PR-checking CI system (physically) which is explained in here: https://cwiki.apache.org/confluence/display/MXNET/Restricted+jobs+and+nodes. What you are mentioning: the Public CIs are all having troubles if they are public accessible. I am not sure how secure the restricted node is. However, the only way I can think of from your end is to downloading all deps in a single machine and run everything there (disconnected from internet). It would bring us the best security we have. Thanks, Qing On 12/17/18, 2:06 PM, "kellen sunderland" wrote: I'm not in favour of publishing artifacts from any Jenkins based systems. There are many ways to bundle artifacts and publish them from an automated system. Why we would use a CI system like Jenkins for this task? Jenkins frequently has security vulnerabilities and is designed to run arbitrary code from the internet. It is a real possibility that an attacker could pivot from any Jenkins based CI system to infect artifacts which would then potentially be pushed to repositories our users would consume. I would consider any system using Jenkins as insecure-by-design, and encourage us to air-gapped any artifact generation (websites, jars, PyPi packages) completely from a system like that. An alternative I could see is a simple Dockerfile (no Jenkins) that builds all artifacts end-to-end and can be run in an automated account well outside our CI account. On Mon, Dec 17, 2018 at 1:53 PM Qing Lan wrote: > Dear community, > > Currently me and Zach are working on the Automated-publish pipeline on > Jenkins which is a pipeline used to publish Maven packages and pip packages > nightly build. We are trying to use NVIDIA deb which could help us to build > different CUDA/CUDNN versions in the publish system. Sheng has provided a > script here: https://github.com/apache/incubator-mxnet/pull/13646. This > provide a very concrete and automatic solution from downloading to > installing on the system. The only scenario we are facing is: It seemed > NVIDIA has a restriction on distributing CUDA. We are not sure if it is > legally-safe for us to use this in public. > > We would be grateful if somebody has a better context on it and help us > out! > > Thanks, > Qing >
Re: [DISCUSS] About the usage of CUDA/CUDNN
I'm not in favour of publishing artifacts from any Jenkins based systems. There are many ways to bundle artifacts and publish them from an automated system. Why we would use a CI system like Jenkins for this task? Jenkins frequently has security vulnerabilities and is designed to run arbitrary code from the internet. It is a real possibility that an attacker could pivot from any Jenkins based CI system to infect artifacts which would then potentially be pushed to repositories our users would consume. I would consider any system using Jenkins as insecure-by-design, and encourage us to air-gapped any artifact generation (websites, jars, PyPi packages) completely from a system like that. An alternative I could see is a simple Dockerfile (no Jenkins) that builds all artifacts end-to-end and can be run in an automated account well outside our CI account. On Mon, Dec 17, 2018 at 1:53 PM Qing Lan wrote: > Dear community, > > Currently me and Zach are working on the Automated-publish pipeline on > Jenkins which is a pipeline used to publish Maven packages and pip packages > nightly build. We are trying to use NVIDIA deb which could help us to build > different CUDA/CUDNN versions in the publish system. Sheng has provided a > script here: https://github.com/apache/incubator-mxnet/pull/13646. This > provide a very concrete and automatic solution from downloading to > installing on the system. The only scenario we are facing is: It seemed > NVIDIA has a restriction on distributing CUDA. We are not sure if it is > legally-safe for us to use this in public. > > We would be grateful if somebody has a better context on it and help us > out! > > Thanks, > Qing >
[DISCUSS] About the usage of CUDA/CUDNN
Dear community, Currently me and Zach are working on the Automated-publish pipeline on Jenkins which is a pipeline used to publish Maven packages and pip packages nightly build. We are trying to use NVIDIA deb which could help us to build different CUDA/CUDNN versions in the publish system. Sheng has provided a script here: https://github.com/apache/incubator-mxnet/pull/13646. This provide a very concrete and automatic solution from downloading to installing on the system. The only scenario we are facing is: It seemed NVIDIA has a restriction on distributing CUDA. We are not sure if it is legally-safe for us to use this in public. We would be grateful if somebody has a better context on it and help us out! Thanks, Qing