Closed
Bug 1306040
Opened 8 years ago
Closed 8 years ago
queue redirecting to cloudfront for EC2 instances
Categories
(Taskcluster :: Services, defect)
Taskcluster
Services
Tracking
(Not tracked)
RESOLVED
INCOMPLETE
People
(Reporter: dustin, Unassigned)
References
Details
Attachments
(1 file)
56.28 KB,
text/plain
|
Details |
Looking just at two hours of logs from the https://taskcluster-public-artifacts.taskcluster.net cloudformation logs, for distinct source IPs, I get about 2.4k distinct EC2 addresses, in both us-east-1 and us-west-2.
From https://github.com/taskcluster/taskcluster-queue/blob/master/src/artifacts.js#L358 it appears that all EC2 requests should be going either directly to the s3 bucket URL or to cloud-mirror.
The region metadata is dated 2016-09-26-16-49-06, and these two hours were the 3:00 hour UTC on the 27th and 28th, specifically logfiles
EXGGJTH3KS8NS.2016-09-27-03.1998f3eb EXGGJTH3KS8NS.2016-09-27-03.f2a2f8ae EXGGJTH3KS8NS.2016-09-28-03.3f5b4a34 EXGGJTH3KS8NS.2016-09-28-03.c91eb9fd
EXGGJTH3KS8NS.2016-09-27-03.234a9959 EXGGJTH3KS8NS.2016-09-27-03.f311188a EXGGJTH3KS8NS.2016-09-28-03.4a18334b EXGGJTH3KS8NS.2016-09-28-03.dd17876b
EXGGJTH3KS8NS.2016-09-27-03.9f5ff394 EXGGJTH3KS8NS.2016-09-27-03.fe2305ac EXGGJTH3KS8NS.2016-09-28-03.8362d369 EXGGJTH3KS8NS.2016-09-27-03.ae0b2232 EXGGJTH3KS8NS.2016-09-28-03.18f6d377 EXGGJTH3KS8NS.2016-09-28-03.9f970f11
so the regions should not have changed since that time.
Reporter | ||
Comment 1•8 years ago
|
||
I can confirm this manually from lamport, which is in us-west-2:
(sandbox) dustin@lamport ~/p/m-c (bug1269443) $ curl -L -v -o /dev/null https://queue.taskcluster.net/v1/task/SDlWxl5OQTOX9u4ZNytuTA/artifacts/public/build/target.reftest.tests.zip
% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
0 0 0 0 0 0 0 0 --:--:-- --:--:-- --:--:-- 0* Trying 54.235.141.80...
* Connected to queue.taskcluster.net (54.235.141.80) port 443 (#0)
* found 173 certificates in /etc/ssl/certs/ca-certificates.crt
* found 692 certificates in /etc/ssl/certs
* ALPN, offering http/1.1
* SSL connection using TLS1.2 / ECDHE_RSA_AES_128_GCM_SHA256
* server certificate verification OK
* server certificate status verification SKIPPED
* common name: auth.taskcluster.net (matched)
* server certificate expiration date OK
* server certificate activation date OK
* certificate public key: RSA
* certificate version: #3
* subject: C=US,ST=California,L=Mountain View,O=Mozilla Corporation,CN=auth.taskcluster.net
* start date: Thu, 17 Mar 2016 00:00:00 GMT
* expire date: Fri, 22 Mar 2019 12:00:00 GMT
* issuer: C=US,O=DigiCert Inc,CN=DigiCert SHA2 Secure Server CA
* compression: NULL
* ALPN, server did not agree to a protocol
0 0 0 0 0 0 0 0 --:--:-- --:--:-- --:--:-- 0> GET /v1/task/SDlWxl5OQTOX9u4ZNytuTA/artifacts/public/build/target.reftest.tests.zip HTTP/1.1
> Host: queue.taskcluster.net
> User-Agent: curl/7.47.0
> Accept: */*
>
< HTTP/1.1 303 See Other
< Server: Cowboy
< Connection: keep-alive
< X-Powered-By: Express
< Strict-Transport-Security: max-age=7776000
< Access-Control-Allow-Origin: *
< Access-Control-Allow-Methods: OPTIONS,GET,HEAD,POST,PUT,DELETE,TRACE,CONNECT
< Access-Control-Request-Method: *
< Access-Control-Allow-Headers: X-Requested-With,Content-Type,Authorization,Accept,Origin
< Location: https://public-artifacts.taskcluster.net/SDlWxl5OQTOX9u4ZNytuTA/0/public/build/target.reftest.tests.zip
< Vary: Accept
< Content-Type: text/plain; charset=utf-8
< Content-Length: 29
< Date: Wed, 28 Sep 2016 16:36:43 GMT
< Via: 1.1 vegur
<
* Ignoring the response-body
{ [29 bytes data]
100 29 100 29 0 0 34 0 --:--:-- --:--:-- --:--:-- 34
* Connection #0 to host queue.taskcluster.net left intact
* Issue another request to this URL: 'https://public-artifacts.taskcluster.net/SDlWxl5OQTOX9u4ZNytuTA/0/public/build/target.reftest.tests.zip'
* Trying 52.84.232.137...
* Connected to public-artifacts.taskcluster.net (52.84.232.137) port 443 (#1)
* found 173 certificates in /etc/ssl/certs/ca-certificates.crt
* found 692 certificates in /etc/ssl/certs
* ALPN, offering http/1.1
* SSL connection using TLS1.2 / ECDHE_RSA_AES_128_GCM_SHA256
* server certificate verification OK
* server certificate status verification SKIPPED
* common name: auth.taskcluster.net (matched)
* server certificate expiration date OK
* server certificate activation date OK
* certificate public key: RSA
* certificate version: #3
* subject: C=US,ST=California,L=Mountain View,O=Mozilla Corporation,CN=auth.taskcluster.net
* start date: Thu, 17 Mar 2016 00:00:00 GMT
* expire date: Fri, 22 Mar 2019 12:00:00 GMT
* issuer: C=US,O=DigiCert Inc,CN=DigiCert SHA2 Secure Server CA
* compression: NULL
* ALPN, server accepted to use http/1.1
> GET /SDlWxl5OQTOX9u4ZNytuTA/0/public/build/target.reftest.tests.zip HTTP/1.1
> Host: public-artifacts.taskcluster.net
> User-Agent: curl/7.47.0
> Accept: */*
>
< HTTP/1.1 200 OK
< Content-Type: application/zip
< Content-Length: 32102631
< Connection: keep-alive
< Date: Wed, 28 Sep 2016 15:12:33 GMT
< Last-Modified: Wed, 28 Sep 2016 03:45:18 GMT
< ETag: "098bdf06f370a606a75637840c4c0b3d"
< x-amz-version-id: jwmF11fp84D3PjyPIbbVORt_v1ecEQvR
< Accept-Ranges: bytes
< Server: AmazonS3
< Age: 5051
< X-Cache: Hit from cloudfront
< Via: 1.1 336f0e6ef9a3462f682d6ca49029b665.cloudfront.net (CloudFront)
< X-Amz-Cf-Id: TdO_6aMqDx0H3Eg1X_bj91Umono9VRWe00bjTKAvj1iL-jO3S-2nZw==
<
{ [16384 bytes data]
100 30.6M 100 30.6M 0 0 18.8M 0 0:00:01 0:00:01 --:--:-- 66.8M
* Connection #1 to host public-artifacts.taskcluster.net left intact
I suspect this is, at the least, an opportunity for cost savings, but may also be related to bug 1305752 and bug 1305768.
Reporter | ||
Updated•8 years ago
|
Summary: queue redirecting to cloudfront for EC2 instances in us-west-2 → queue redirecting to cloudfront for EC2 instances
Reporter | ||
Comment 2•8 years ago
|
||
The 2438 distinct IPs I found, tagged with region. I downloaded and unzipped the CF logs, then
cut -d' ' -f 5 * | sort -u > ips
then ran
import requests
import json
from IPy import *
ips = [IP(ip.strip()) for ip in open("ips")]
ranges = requests.get('https://ip-ranges.amazonaws.com/ip-ranges.json').json()
ranges = [(IP(pfx['ip_prefix']), pfx['region']) for pfx in ranges['prefixes']]
for ip in ips:
region = 'unknown'
for pfx, rgn in ranges:
if ip in pfx:
region = rgn
break
print region, ip
Comment 3•8 years ago
|
||
I submitted https://github.com/taskcluster/taskcluster-queue/pull/120 to help figure out what the queue's view of the requester is
Comment 4•8 years ago
|
||
PR fixing the issue with us-west-2:
https://github.com/taskcluster/taskcluster-queue/pull/121
Comment 5•8 years ago
|
||
I merged above PR that should fix us-west-2 and give us some more debug statements.
I suspect we ought to refactor, but really, we want to move region resolution into cloud-mirror.
Comment hidden (typo) |
Comment hidden (typo) |
Comment 8•8 years ago
|
||
Dustin says I can close this.
Status: NEW → RESOLVED
Closed: 8 years ago
Resolution: --- → INCOMPLETE
Assignee | ||
Updated•6 years ago
|
Component: Queue → Services
You need to log in
before you can comment on or make changes to this bug.
Description
•