Skip to content

Fixed writing clean swap file broken in 4310f8b#155

Merged
rousskov merged 1 commit intosquid-cache:masterfrom
measurement-factory:SQUID-357-empty-ufs-cache-index-after-restart
Feb 22, 2018
Merged

Fixed writing clean swap file broken in 4310f8b#155
rousskov merged 1 commit intosquid-cache:masterfrom
measurement-factory:SQUID-357-empty-ufs-cache-index-after-restart

Conversation

@eduard-bagdasaryan
Copy link
Contributor

No description provided.

@eduard-bagdasaryan
Copy link
Contributor Author

Here is squid-users related discussion.

Copy link
Contributor

@rousskov rousskov left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM.

@rousskov
Copy link
Contributor

Jenkins build failed on d-centos-7, but the failures are completely unrelated to this (trivial) PR change:

src/security/KeyData.cc:57:53: error: 'gnutls_pcert_export_x509' was not declared in this scope
src/security/KeyData.cc:148:50: error: 'gnutls_privkey_export_x509' was not declared in this scope

I am guessing that Jenkins build environment itself has changed, possibly exposing a previously unknown GnuTLS support bug introduced by rogue 51e09c0 commit (PR #81). Comparing the old successful PR #81 build log with this PR failed build log, I see a new -I preprocessor option added, but I do not know who/what added that option and whether that addition caused the build failure:

old: configure: BUILD Tools C++ FLAGS:  -g -O2 -std=c++11
new: configure: BUILD Tools C++ FLAGS:  -I/usr/include/p11-kit-1   -g -O2 -std=c++11

@squid-cache/devops, if you know what happened and can easily undo that d-centos-7 change, please undo (at least for now). @yadij, if this compilation error is legitimate, please fix the bug. All d-centos-7 tests will fail until this is resolved.

Meanwhile, I plan to ignore these unrelated failures and merge this trivial PR that fixes a high-impact bug.

@yadij
Copy link
Contributor

yadij commented Feb 22, 2018

All of the build nodes have had upgrades a few days go instigated by some recent security issues related to SSH and Ruby which affected Jenkins indirectly.

I am looking into it right now.

Okay from me on ignoring this nodes result for this PR.

@rousskov
Copy link
Contributor

All of the build nodes have had upgrades a few days go instigated by some recent security issues related to SSH and Ruby which affected Jenkins indirectly.

IMHO, build node security should be based primarily on isolation rather than (frequent) updates -- an ideal build node does not communicate with the outside world. Any upgrades should be done by building an updated clone and validating that it passes build tests of major official branches. I do not know how far we are from that ideal, but we should be aiming for it.

@rousskov rousskov merged commit 8080123 into squid-cache:master Feb 22, 2018
@rousskov rousskov deleted the SQUID-357-empty-ufs-cache-index-after-restart branch February 22, 2018 15:38
squidadm pushed a commit to squidadm/squid that referenced this pull request Feb 23, 2018
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants