Bug 1381433
| Summary: | Need access to machines which run smoke | ||||||
|---|---|---|---|---|---|---|---|
| Product: | [Community] GlusterFS | Reporter: | Pranith Kumar K <pkarampu> | ||||
| Component: | project-infrastructure | Assignee: | Nigel Babu <nigelb> | ||||
| Status: | CLOSED CURRENTRELEASE | QA Contact: | |||||
| Severity: | unspecified | Docs Contact: | |||||
| Priority: | unspecified | ||||||
| Version: | mainline | CC: | bugs, gluster-infra, misc, pkarampu | ||||
| Target Milestone: | --- | Keywords: | Triaged | ||||
| Target Release: | --- | ||||||
| Hardware: | Unspecified | ||||||
| OS: | Unspecified | ||||||
| Whiteboard: | |||||||
| Fixed In Version: | Doc Type: | If docs needed, set a value | |||||
| Doc Text: | Story Points: | --- | |||||
| Clone Of: | Environment: | ||||||
| Last Closed: | 2016-10-19 04:15:53 UTC | Type: | Bug | ||||
| Regression: | --- | Mount Type: | --- | ||||
| Documentation: | --- | CRM: | |||||
| Verified Versions: | Category: | --- | |||||
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |||||
| Cloudforms Team: | --- | Target Upstream Version: | |||||
| Embargoed: | |||||||
| Attachments: |
|
||||||
I would like to get a bit more justification, if only for tracking. Oops, sorry, the context is that we are trying to find RC for why smoke is hanging. To find that out we need a machine to run smoke in a loop and see where it is hanging. To reserve one of the machines for this purpose Nigel asked me to raise a bug with public-key. You should be all set on slave33.cloud.gluster.org. Please ssh in as jenkins.gluster.org. Remember to delete any temporary files you create and undo any script changes you make. Apologies, it's slave34, not slave33. I'm guessing you need more time since the usual 1 week time ends tomorrow. I'll check back in next Tuesday/Wednesday. Nigel,
There is some peculiar behavior, like Milind was mentioning yesterday, it happened only the first time and never after that. Wonder what the trigger is. I think the best way forward for this hang is to improve our smoke.sh to take statedumps before killing the hung mount, which would give us more clue about why it hung. Let me ping you on IRC about the necessary things we can do! Take the machine back.
Pranith
Created attachment 1211787 [details]
Collect statedumps before bricks/mounts are killed in smoke.sh
Attaching the modified smoke.sh, Can we deploy this on the build machines which run smoke? We can probably get a review too before we deploy it on all the machines? I am not sure about the practice.
Hrm, have you thought of changing the gluster commit, rebuilding, and trying smoke again? That seems to be difference between "first time" and later attempts. Did that once, but didn't help. Since we anyway need smoke.sh to get better, that seemed like a better approach for now. I will be submitting a pull request today sometime. Let's see. I'll close this bug and put the machine back in the pool, then. |
Description of problem: hi Nigel, This is my public key: 11:17:50 :) ⚡ cat ~/.ssh/id_rsa.pub ssh-rsa AAAAB3NzaC1yc2EAAAADAQABAAABAQDkduuGBq++zm/JKYVUcfM6YOqzYp2Dj0ag3OvlkFTXyNZ1QVOoEWuH9MAeF/MlHd14nLvFKSdpI+qr+faY+Wtyt/Za09YnizyMBuEo9hIw307EwynOdfAO8N/PKLAvtsNQ7Xk3UHUfHrvVuJr5qZFs1sWNau67/DBxd3bUO/FUl3FZoZqWg3/qsG8ZTCVEPc4N0qY9xiDFxgDh81lmK8t24S8d9RfMrKtpPbSe75HW1CxqM6AGLpQtDscIydGqmRYYcYSn9box4T3erbVxNpcpSlk6K1akMJhbuNoEbDfD7n4t8X/BLj/h3gJIUTlrXnpPj+hluiHDmeBlhu7a7ctd pk.eng.blr.redhat.com Version-Release number of selected component (if applicable): How reproducible: Steps to Reproduce: 1. 2. 3. Actual results: Expected results: Additional info: