update-<timestamp> directories being left behind
Bug #1220269 reported by
Sidnei da Silva
This bug affects 1 person
| Affects | Status | Importance | Assigned to | Milestone | |
|---|---|---|---|---|---|
| juju-core |
Fix Released
|
High
|
Sidnei da Silva | ||
Bug Description
Apparently the collectOrphans() might be failing silently. No visible signs in the log of what might have failed. One instance had:
46G /var/lib/
This instance is auto-upgraded every 20min as part of a CI environment, and the charm has a 260Mb tarball payload. There were 60+ update-<timestamp> dirs left behind, most with ~264Mb each.
Related branches
lp:~sidnei/juju-core/fix-cleanup-orphans
- Juju Engineering: Pending requested
-
Diff: 224 lines (+72/-13)3 files modifiedworker/uniter/charm/deployer.go (+23/-7)
worker/uniter/charm/deployer_test.go (+35/-6)
worker/uniter/charm/export_test.go (+14/-0)
| description: | updated |
| tags: | added: canonical-webops |
Revision history for this message
| JuanJo Ciarlante (jjo) wrote | #1 |
| Changed in juju-core: | |
| assignee: | nobody → Sidnei da Silva (sidnei) |
| importance: | Undecided → High |
| status: | New → In Progress |
| Changed in juju-core: | |
| status: | In Progress → Fix Committed |
| Changed in juju-core: | |
| milestone: | none → 1.15.0 |
| Changed in juju-core: | |
| status: | Fix Committed → Fix Released |
To post a comment you must log in.
Details from a (different charm's) unit, showing the same issue:
root@foo-
4.0K current
787M update-
264M update-
265M update-
4.5M update-
264M update-
265M update-
264M update-
265M update-
264M update-
265M update-
264M update-
265M update-
264M update-
265M update-
264M update-
265M update-
264M update-
265M update-
264M update-
265M update-
264M update-
265M update-
264M update-
265M update-
231M update-
6.7G total
* can clearly see the one-per-20min update frequency (it's cron-ed to be so)
* directory size matches charm size (it carries a ~260M .tgz code payload)