ucbrise / cirrus Goto Github PK

The code references this s3 bucket without much explaination: cirrus-criteo-kaggle-19b-random. Could I get any information on what's inside this bucket? What should I put in this bucket?

This is the Gdb output:

Thread 4 "test_libsvm_ite" received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0x7ffff71ad700 (LWP 9489)]
0x00007ffff676e448 in internal_getent (stream=stream@entry=0x7ffff5c4f720, result=result@entry=0x7ffff71ac420, buffer=buffer@entry=0x7ffff71ac830 "s3-r-w.us-west-2.amazonaws.com", 
    buflen=buflen@entry=1024, errnop=errnop@entry=0x7ffff71ad6b0, herrnop=herrnop@entry=0x7ffff71ad6e8, af=0, flags=0) at nss_files/files-XXX.c:216
216	nss_files/files-XXX.c: No such file or directory.
(gdb) bt
#0  0x00007ffff676e448 in internal_getent (stream=stream@entry=0x7ffff5c4f720, result=result@entry=0x7ffff71ac420, buffer=buffer@entry=0x7ffff71ac830 "s3-r-w.us-west-2.amazonaws.com", 
    buflen=buflen@entry=1024, errnop=errnop@entry=0x7ffff71ad6b0, herrnop=herrnop@entry=0x7ffff71ad6e8, af=0, flags=0) at nss_files/files-XXX.c:216
#1  0x00007ffff676f8bd in _nss_files_gethostbyname4_r (name=0x7ffff00b42a0 "criteo-kaggle-libsvm-train.s3.us-west-2.amazonaws.com", pat=0x7ffff71ac5a8, 
    buffer=0x7ffff71ac830 "s3-r-w.us-west-2.amazonaws.com", buflen=1024, errnop=0x7ffff71ad6b0, herrnop=0x7ffff71ad6e8, ttlp=0x0) at nss_files/files-hosts.c:416
#2  0x00000000009bd5a5 in gaih_inet.constprop ()
#3  0x00000000009bf534 in getaddrinfo ()
#4  0x00000000006a8257 in Curl_getaddrinfo_ex ()
#5  0x00000000006a9b3f in getaddrinfo_thread ()
#6  0x00000000006d077b in curl_thread_create_thunk ()
#7  0x0000000000404a3b in start_thread (arg=0x7ffff71ad700) at pthread_create.c:463
#8  0x00000000009c37ef in clone ()

It seems to be related to this issue: https://bugs.launchpad.net/ubuntu/+source/glibc/+bug/1613067

By the way, is there any reason to statically linking here?

An extra null pointer check is not needed in the destructor for the class “CircularBuffer”.

Would you like to wrap any pointer variables with the class template “std::unique_ptr”?

Here are some adjustments to the interface that would make this more useable for other researchers who want to run or benchmark this. (In the meantime, hopefully this helps anyone else who wants to launch this. I didn't manage to fully launch it.)

• Allow AWS region to passed as a config value instead of hardcoding the AWS region to be us-west-2 (both in the parameter server and in the Python front-end here and here)
• Don't hard-code the path to the worker function bundle as a global path that won't match the path for other users
• When front-end is sshing to another machine to start the parameter server, use the ssh key file (with ssh -i ...) provided in the config
• Don't hardcode VPC settings but make the user specify them
• Describe briefly what this is and how to run it in the README -- it's not clear how to run the frontend. (It's located in python/frontend/cirrus, and the examples directory gives some guidance as to how to run the frontend.)
• Link or cite the paper Cirrus: a Serverless Framework for End-to-end ML Workflows in the README

I find that include guards like “MODEL_H_” and “TASKS_H_” are too short for the safe reuse of your header files (when they belong to an application programming interface).

Is there any documentation on how to deploy cirrus to aws? It seems I need to create a vm to launch the parameter server? How does the lambda server connect to the parameter server?

--2020-12-22 09:46:43-- (try:15) http://people.redhat.com/~dhowells/keyutils/keyutils-1.5.10.tar.bz2
Connecting to people.redhat.com (people.redhat.com)|209.132.183.19|:80... connected.
HTTP request sent, awaiting response... Read error (Connection reset by peer) in headers.
Retrying.

I would like to point out that identifiers like “_INPUT_H_” and “_MFMODEL_H_” do eventually not fit to the expected naming convention of the C++ language standard.
Would you like to adjust your selection for unique names?