Switch to vault.centos.org

[TimNN]

Centos 5 is EOL and is no longer available from the usual mirrors, so switch to http://vault.centos.org/

r? @alexcrichton

(I changed the sed commands to apply to all *.repo files, since there were some bad urls in the other ones as well, I think).

[alexcrichton]

@bors: r+ p=10000

[bors]

📌 Commit 802c11c has been approved by alexcrichton

[bors]

⌛ Testing commit 802c11c with merge 5677ac8...

[bors]

💔 Test failed - status-travis

[frewsxcv]

@bors retry

[bors]

⌛ Testing commit 802c11c with merge 7a968d1...

[frewsxcv]

@bors retry

[frewsxcv]

@bors r+

[bors]

📌 Commit 6dc8869 has been approved by frewsxcv

[bors]

⌛ Testing commit 6dc8869 with merge 2b79138...

[aidanhs]

Just to record this somewhere, dnschecker.org currently lists the following IPs (when deduplicated) for vault.centos.org:

88.208.217.170
107.158.252.35
108.61.16.227
109.72.83.61
160.10.26.25

Of these five (which I suspect are a tiny subset), two would definitely fail:

$ for f in $(cat y); do curl -sf http://$f/centos/5/os/x86_64/repodata/ >/dev/null || echo bad $f; done
bad 108.61.16.227
bad 160.10.26.25

Also, a relevant post from centos-devel about slow mirror updates last month: https://lists.centos.org/pipermail/centos-devel/2017-March/015708.html

[TimNN]

@aidanhs: I would assume that vault.centos.org doesn't use the official mirrors of the non-eol centos releases.

I believe that the ip's we are seeing for vault.centos.org are all maintained by the centos project -- http://vault.centos.org/readme.txt lists some "real" mirrors.

[frewsxcv]

@bors delegate=TimNN

Giving you review privileges in case Travis fails so you can 'bors retry'

[bors]

✌️ @TimNN can now approve this pull request

[aidanhs]

@TimNN yeah good point, I stumbled across that page just after posting the comment above but they currently look like they need to catch up too. I'd lean towards working on a non-centos5 strategy after this has merged rather than moving over to a real mirror.

[bors]

💔 Test failed - status-travis

[mrhota]

😞 build merely took too long and was terminated

[frewsxcv]

@bors retry

[TimNN]

@bors retry

• ar.exe failed

[bors]

⌛ Testing commit 6dc8869 with merge a9214df...

[bors]

💔 Test failed - status-travis

[TimNN]

@bors retry

• Spurious failures in sccache #40240

[bors]

⌛ Testing commit 6dc8869 with merge c417350...

[TimNN]

(By the way, just to establish the link, the ar.exe failure is #40546 (Don't want to edit the previous comment or bo_s will pick up the retry again))

[bors]

💔 Test failed - status-travis

[frewsxcv]

@bors retry

[frewsxcv]

It seems to consistently be dist-arm-linux causing issues here.

Here's the build log for the last fail for that machine:

https://api.travis-ci.org/jobs/218400741/log.txt?deansi=true

The build successfully finished at 9:01AM EST, but then start doing cleanup and caching and reached the three hour mark at 9:06AM EST.

[aidanhs]

For people wanting to read the log, build starts at Tue, 04 Apr 2017 10:07:59 GMT ("1491300479230695090") and the before_cache steps finish at Tue, 04 Apr 2017 13:06:17 GMT ("1491311177686306522").

[bors]

⌛ Testing commit 6dc8869 with merge a5cce63...

[TimNN]

@bors retry

[bors]

⌛ Testing commit 6dc8869 with merge 89db846...

[TimNN]

@bors retry

[bors]

⌛ Testing commit 6dc8869 with merge 7a966a4...

[TimNN]

@bors retry

[bors]

⌛ Testing commit 6dc8869 with merge 2564711...