~tcuthbert/ubuntu-repository-cache/squid-mem-max

From Revision 330 to 311

expand all

collapse all

Rev		Summary	Authors	Date	Diff	Files
330		Add 'Accept-Encoding' in nrpe and health checks Add 'Accept-Encoding' in nrpe and health checks Upstream archives use mod_deflate which sets 'Vary: Accept-Encoding' on responses, so check_http should set that header. This also reduces the logging spam from squid: | 2021/03/10 00:15:00 kid1| varyEvaluateMatch: Oops. Not a Vary match on second attempt, 'http://archive.ubuntu.com/ubuntu/pool/' 'accept-encoding="*"' | 2021/03/10 00:15:00 kid1| clientProcessHit: Vary object loop! We also hope to reduce squid memory usage (mem_node) as discovered by tcuthbert. Reviewed-on: https://code.launchpad.net/~hloeung/ubuntu-repository-cache/add-accept-encoding-headers/+merge/399399 Reviewed-by: Thomas Cuthbert <thomas.cuthbert@canonical.com>	"Haw Loeung"	3 years ago
329		Fix issue with units stuck in maintenance - LP:1918199 Fix issue with units stuck in maintenance - LP:1918199 Reviewed-on: https://code.launchpad.net/~hloeung/ubuntu-repository-cache/units-stuck-in-maintenance/+merge/399324 Reviewed-by: Barry Price <barry.price@canonical.com>	"Haw Loeung"	3 years ago
328		Don't cycle out apache2 processes by default - LP:1918211 Don't cycle out apache2 processes by default - LP:1918211 This reduces the frequent and brief outages. Reviewed-on: https://code.launchpad.net/~hloeung/ubuntu-repository-cache/no-apache2-mpm-MaxConnectionsPerChild/+merge/399336 Reviewed-by: Barry Price <barry.price@canonical.com>	"Haw Loeung"	3 years ago
327		Make health-check script/endpoint more resilient Make health-check script/endpoint more resilient u-r-c units are usually deployed in public clouds with the sync-host set to the main Ubuntu archive servers. We can't control the network connectivity/conditions so update health-check script to retry up to 3 times on failures. Reviewed-on: https://code.launchpad.net/~hloeung/ubuntu-repository-cache/health-check-retries-on-failure/+merge/399257 Reviewed-by: Benjamin Allot <benjamin.allot@canonical.com> Reviewed-by: Barry Price <barry.price@canonical.com>	"Haw Loeung"	3 years ago
326		Don't set health-check-disabled flag in apache.create_metada... Don't set health-check-disabled flag in apache.create_metadata_site() On non-leader units, metadata sync calls service.pause() which then calls this. This causes a brief period where the unit is disabled and not serving requests. It's useful for initial unit provisioning but not so on a fully functional unit where metadata points to an existing copy on disk. NOTE: The setting of health-check-disabled flag was introduced in r309 (MP:398100). Reviewed-on: https://code.launchpad.net/~hloeung/ubuntu-repository-cache/dont-disable-unit-on-failover-or-pause/+merge/399254 Reviewed-by: Paul Collins <paul.collins@canonical.com>	"Haw Loeung"	3 years ago
325		[trivial] Minor tidy up [trivial] Minor tidy up	Haw Loeung	3 years ago
324		Log overrall check result and URLs for errors Log overrall check result and URLs for errors Reviewed-on: https://code.launchpad.net/~hloeung/ubuntu-repository-cache/health-check-logging-improvements/+merge/399130 Reviewed-by: Colin Misare <colin.misare@canonical.com>	"Haw Loeung"	3 years ago
323		Avoid holding up connection/request based on sending metrics... Avoid holding up connection/request based on sending metrics and flush buffers - LP:1917317 Reviewed-on: https://code.launchpad.net/~hloeung/ubuntu-repository-cache/health-check-flush-output-buffers/+merge/399129 Reviewed-by: David Lawson <dave.lawson@canonical.com>	"Haw Loeung"	3 years ago
322		Send metrics at the very end - LP:1917317 Send metrics at the very end - LP:1917317 Reviewed-on: https://code.launchpad.net/~hloeung/ubuntu-repository-cache/health-check-send-metrics-at-end/+merge/399128 Reviewed-by: Alexandre Gomes <alexandre.gomes@canonical.com>	"Haw Loeung"	3 years ago
321		Include syncing dist-upgrade-all metadata - LP:1765934 Include syncing dist-upgrade-all metadata - LP:1765934 Reviewed-on: https://code.launchpad.net/~hloeung/ubuntu-repository-cache/sync-dist-upgrade-all-metadata/+merge/398938 Reviewed-by: Colin Misare <colin.misare@canonical.com>	"Haw Loeung"	3 years ago
320		We're already logging, so may as well log to the right place... We're already logging, so may as well log to the right place - LP:1917385 Reviewed-on: https://code.launchpad.net/~hloeung/ubuntu-repository-cache/fix-squid-access-log-location/+merge/399035 Reviewed-by: Paul Collins <paul.collins@canonical.com>	"Haw Loeung"	3 years ago
319		Add socket timeout for sending metrics to influx Add socket timeout for sending metrics to influx Reviewed-on: https://code.launchpad.net/~hloeung/ubuntu-repository-cache/add-send-to-influx-socket-timeout/+merge/398931 Reviewed-by: Paul Collins <paul.collins@canonical.com>	"Haw Loeung"	3 years ago
318		Add apache2 cgi script timeout Add apache2 cgi script timeout Otherwise, value of `Timeout` is used and that defaults to 60secs which is too long. We have two check points (metadata and pool) with a timeout in the health-check cgi script to 10secs for each so 20secs should be enough for both. Reviewed-on: https://code.launchpad.net/~hloeung/ubuntu-repository-cache/apache2-cgi-script-timeout/+merge/398920 Reviewed-by: Benjamin Allot <benjamin.allot@canonical.com>	"Haw Loeung"	3 years ago
317		Bump metadata sync timeout Bump metadata sync timeout As seem in AWS us-east-1, it can take over an hour to sync the metadata (which is now 26GiB) from another unit in the same region. Reviewed-on: https://code.launchpad.net/~hloeung/ubuntu-repository-cache/bump-metadata-sync-timeout/+merge/398848 Reviewed-by: Max Horlanchuk <max.horlanchuk@canonical.com>	"Haw Loeung"	3 years ago
316		Only reload/graceful apache2 when required - LP:1916084 Only reload/graceful apache2 when required - LP:1916084 Reviewed-on: https://code.launchpad.net/~hloeung/ubuntu-repository-cache/only-reload-apache2-when-needed/+merge/398325 Reviewed-by: Joel Sing <joel.sing@canonical.com>	"Haw Loeung"	3 years ago
315		unittests now uses hookenv.Config() so ensure .juju-persiste... unittests now uses hookenv.Config() so ensure .juju-persistent-config isn't included or is properly cleaned Reviewed-on: https://code.launchpad.net/~hloeung/ubuntu-repository-cache/cleanup-for-charm-build/+merge/398499 Reviewed-by: Colin Misare <colin.misare@canonical.com>	"Haw Loeung"	3 years ago
314		Don't serve health-check files if cgid module isn't load Don't serve health-check files if cgid module isn't load Reviewed-on: https://code.launchpad.net/~hloeung/ubuntu-repository-cache/missing-apache2-cgi-mod-2/+merge/398492 Reviewed-by: Benjamin Allot <benjamin.allot@canonical.com>	"Haw Loeung"	3 years ago
313		Enable cgi module on install otherwise health check succeeds... Enable cgi module on install otherwise health check succeeds with output from the script; also we want cgid for multi-threaded apache2 Reviewed-on: https://code.launchpad.net/~hloeung/ubuntu-repository-cache/missing-apache2-cgi-mod/+merge/398421 Reviewed-by: Paul Collins <paul.collins@canonical.com>	"Haw Loeung"	3 years ago
312		Don't call apache.start() in service.unpause() Don't call apache.start() in service.unpause() service.unpause() calls render_configs() which calls apache.render_configs(). Within that, it will call apache.start() so it is not needed here. It also handles config changes requiring graceful reload better. Reviewed-on: https://code.launchpad.net/~hloeung/ubuntu-repository-cache/only-reload-apache2-when-needed-4/+merge/398419 Reviewed-by: Paul Collins <paul.collins@canonical.com>	"Haw Loeung"	3 years ago
311		Only reload/graceful apache2 when required - LP:1916084 Only reload/graceful apache2 when required - LP:1916084 Reviewed-on: https://code.launchpad.net/~hloeung/ubuntu-repository-cache/only-reload-apache2-when-needed-3/+merge/398328 Reviewed-by: Joel Sing <joel.sing@canonical.com>	"Haw Loeung"	3 years ago

• « Newer
• Older »

Loggerhead is a web-based interface for Breezy
Version: 2.0.1