Github user jacksontj closed the pull request at:
https://github.com/apache/trafficserver/pull/526
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the featu
Github user jacksontj commented on the pull request:
https://github.com/apache/trafficserver/pull/526#issuecomment-199878278
I don't want to merge this PR in, its a definite hack. I'm working on a
better fix, it'll just take some time.
---
If your project is set up for it, you can re
Github user zwoop commented on the pull request:
https://github.com/apache/trafficserver/pull/526#issuecomment-199871069
@SolidWallOfCode to review.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does no
Github user jacksontj commented on the pull request:
https://github.com/apache/trafficserver/pull/526#issuecomment-197055859
From what I'm seeing 1) should be completely doable-- I'll take a crack at
it :)
---
If your project is set up for it, you can reply to this email and have you
Github user zwoop commented on the pull request:
https://github.com/apache/trafficserver/pull/526#issuecomment-197047822
Not sure I understand. I'm saying, evict before it gets full, so normal
eviction happens when there's 10% of less available, before inserting new
stuff. That gives
Github user jacksontj commented on the pull request:
https://github.com/apache/trafficserver/pull/526#issuecomment-197011767
I don't think we can ever gaurantee that there will always be space
available, and since we already evict when we need space, we should just keep
doing that IMO
Github user zwoop commented on the pull request:
https://github.com/apache/trafficserver/pull/526#issuecomment-196975636
Maybe it needs to evict things more aggressively, such that it always has
~10% head room for these sort of events ?
---
If your project is set up for it, you can r
Github user jacksontj commented on the pull request:
https://github.com/apache/trafficserver/pull/526#issuecomment-196921878
As I'm thinking through this some more, I'm not certain this is a great
solution either. The particular case that I reproduced was that hostdb was full
when ret
Github user jacksontj commented on the pull request:
https://github.com/apache/trafficserver/pull/526#issuecomment-196907474
This is a different patch to the same issue. Instead of scheduling it for a
retry, we can just return the non-mmaped HostDBInfo we have, so everything else
can
Github user jacksontj commented on the pull request:
https://github.com/apache/trafficserver/pull/526#issuecomment-196904843
I'm not sure this is the correct fix, in this case where we have no space
we probably should just get lookup_done to return a non-persistant version of
the reco
GitHub user jacksontj opened a pull request:
https://github.com/apache/trafficserver/pull/526
Fix for TS-4276
In the event that `lookup_done` returns a NULL, we'll reschedule the lookup
in the future instead of dumping core.
You can merge this pull request into a Git repository by
11 matches
Mail list logo
