Bug #13472
Replace www.centos.org in htpdate pools
100%
Description
It appears to fail, and so time syncing sometimes fails too in Tails when this host is picked up in the pool.
Let replace it by something like getfedora.org.
Subtasks
Related issues
| Related to Tails - Bug #10495: The 'the time has synced' step is fragile | In Progress | 2015-11-06 | |
| Related to Tails - Bug #11562: Monitor servers from the htpdate pools | Confirmed | 2016-07-14 |
History
#1 Updated by bertagaz 2017-07-15 14:44:06
- related to Bug #10495: The 'the time has synced' step is fragile added
#2 Updated by bertagaz 2017-07-15 14:44:23
- related to Bug #11562: Monitor servers from the htpdate pools added
#3 Updated by bertagaz 2017-07-15 14:47:47
- Status changed from Confirmed to In Progress
- Priority changed from Normal to Elevated
- % Done changed from 0 to 50
- Feature Branch set to bugfix/13472-replace-centos-in-htpdate-pool
Let see how it goes in Jenkins.
#4 Updated by intrigeri 2017-07-25 09:44:38
Hi! I’d like to review (and hopefully merge) this in the next few days. Assuming my code review passes and Jenkins is happy, shall I proceed or is there anything else you want to do first?
#5 Updated by bertagaz 2017-07-27 10:03:38
intrigeri wrote:
> Hi! I’d like to review (and hopefully merge) this in the next few days. Assuming my code review passes and Jenkins is happy, shall I proceed or is there anything else you want to do first?
Nop, not yet. I still see “time has synced” failures in this branch too, so I’m not sure it’s really fixing something. I’m debugging $HOME. My next move is to push a patch that stores the htpdate logs on failure of this step I think.
#6 Updated by intrigeri 2017-07-27 10:18:20
> I still see “time has synced” failures in this branch too, so I’m not sure it’s really fixing something.
IMO, the fact that it’s not enough to fix the broader problem (i.e. time sync’ing sometimes fails) should not prevent us from applying a trivial fix to some part of the problem (which is what this ticket is about). No?
#7 Updated by bertagaz 2017-07-27 15:50:32
intrigeri wrote:
> IMO, the fact that it’s not enough to fix the broader problem (i.e. time sync’ing sometimes fails) should not prevent us from applying a trivial fix to some part of the problem (which is what this ticket is about). No?
Yeah, but I’m not so sure about centos.org anymore, I’ve also seen it working sometimes, so I wonder if that was not just a Tor circuit error. Need to check a bit more.
#8 Updated by intrigeri 2017-07-27 16:03:23
> Yeah, but I’m not so sure about centos.org anymore, I’ve also seen it working sometimes, so I wonder if that was not just a Tor circuit error. Need to check a bit more.
I’ve tried it today (in Tor Browser) and I could not connect. So at the very least, it seems too unreliable to be used in our HTP pool. The fact that it works “sometimes” doesn’t really matter :)
#9 Updated by bertagaz 2017-08-03 11:08:34
- Status changed from In Progress to Fix committed
- Assignee deleted (
bertagaz) - % Done changed from 50 to 100
- QA Check set to Pass
intrigeri wrote:
> I’ve tried it today (in Tor Browser) and I could not connect. So at the very least, it seems too unreliable to be used in our HTP pool. The fact that it works “sometimes” doesn’t really matter :)
Meanwhile more of my own tests showed too that centos.org was not reliable enough, so I’ve merged this branch in stable and devel.
But time syncing is still failing quite often right now. Let track the rest of the work on that with Bug #10495.
#10 Updated by intrigeri 2017-08-03 13:02:46
> Meanwhile more of my own tests showed too that centos.org was not reliable enough, so I’ve merged this branch in stable and devel.
Thank you for working on this.
Nitpicking: your change breaks the lexical sorting of HTP_POOL_NEUTRAL. One more reason not to bypass our usual review process. But sure, yeah, whatever :)
#11 Updated by bertagaz 2017-08-09 12:38:26
- Status changed from Fix committed to Resolved