Bug 614303
| Summary: | ssh dump doesn't work with 'cp' core_collector | ||||||
|---|---|---|---|---|---|---|---|
| Product: | Red Hat Enterprise Linux 6 | Reporter: | Cong Wang <amwang> | ||||
| Component: | kexec-tools | Assignee: | Neil Horman <nhorman> | ||||
| Status: | CLOSED CURRENTRELEASE | QA Contact: | Chao Ye <cye> | ||||
| Severity: | high | Docs Contact: | |||||
| Priority: | high | ||||||
| Version: | 6.0 | CC: | amwang, cye, nhorman, phan, qcai, rkhan, robotsari, syeghiay | ||||
| Target Milestone: | rc | ||||||
| Target Release: | --- | ||||||
| Hardware: | All | ||||||
| OS: | Linux | ||||||
| Whiteboard: | |||||||
| Fixed In Version: | kexec-tools-2_0_0-117_el6 | Doc Type: | Bug Fix | ||||
| Doc Text: | Story Points: | --- | |||||
| Clone Of: | Environment: | ||||||
| Last Closed: | 2010-11-11 14:45:16 UTC | Type: | --- | ||||
| Regression: | --- | Mount Type: | --- | ||||
| Documentation: | --- | CRM: | |||||
| Verified Versions: | Category: | --- | |||||
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |||||
| Cloudforms Team: | --- | Target Upstream Version: | |||||
| Embargoed: | |||||||
| Attachments: |
|
||||||
|
Description
Cong Wang
2010-07-14 04:34:46 UTC
Created attachment 431807 [details]
patch to ignore core_collector if its not makedumpfile
Amerigo, can you please try this patch for me (I don't have a free system to test on at the moemnt). It basically just ignores core_collector if its not makedumpfile, which I think is the only sane thing to do when using scp.
It works, but I replaced the 'grep' with bash variable substitution, I think this is simpler. I will check in the modified version of your patch. Thanks. (In reply to comment #2) > Created an attachment (id=431807) [details] > patch to ignore core_collector if its not makedumpfile > > Amerigo, can you please try this patch for me (I don't have a free system to > test on at the moemnt). It basically just ignores core_collector if its not > makedumpfile, which I think is the only sane thing to do when using scp. Hi, I am wondering if this is the best solution. I invoke a custom script as my core_collector, to account for the different ways of invoking makedumpfile on baremetal vs. dom0 kernels (I provide myself some options and extensions on top of the basic makedumpfile). My script does call makedumpfile and will write to stdout in this place but the name of my script is obviously not makedumpfile. So I worry that grepping for makedumpfile straight in the core_collector name is too restrictive and could affect people calling their own scripts. I will try to think of a better way to get around this. I guess I can rename my script such that "makedumpfile" is contained within it, but that does become bulky and requires updating in many locations where it is being used. Anyways just a thought. I don't know if other users would be running a custom script as their core_collector. (In reply to comment #4) > I am wondering if this is the best solution. I invoke a custom script as my > core_collector, to account for the different ways of invoking makedumpfile on > baremetal vs. dom0 kernels (I provide myself some options and extensions on top > of the basic makedumpfile). My script does call makedumpfile and will write to > stdout in this place but the name of my script is obviously not makedumpfile. > So I worry that grepping for makedumpfile straight in the core_collector name > is too restrictive and could affect people calling their own scripts. > This is reasonable, but I am not sure. The interface of CORE_COLLECTOR is pretty limited, it requires sometimes to copy to stdout (then it only accepts one argument), sometimes to copy to another core file (then it accepts two arguments). The best solution I can think of is to invent different CORE_COLLECTOR for different dump, for example SSH_CORE_COLLECTOR for ssh dump, and document the requirements. (In reply to comment #5) > This is reasonable, but I am not sure. The interface of CORE_COLLECTOR is > pretty limited, it requires sometimes to copy to stdout (then it only accepts > one argument), sometimes to copy to another core file (then it accepts two > arguments). > > The best solution I can think of is to invent different CORE_COLLECTOR for > different dump, for example SSH_CORE_COLLECTOR for ssh dump, and document the > requirements. I think that is a good idea. It would certainly clear up a lot of usage issues to have more specific options rather than one catchall option with a bunch of different ways of using it. Document is also nice. I have trouble using kdump and crash sometimes because features are added without being documented. grr :) kdump.conf: net root.65.161 core_collector cp default shell Verified with -116.el6: ============================================================ ...... Saving to remote location root.65.161 1+0 records in 1+0 records out Free memory/Total memory (free %): 215904 / 249524 ( 86.5263 ) BusyBox v1.15.1 (2010-06-01 08:04:20 EDT) multi-call binary Usage: cp [OPTIONS] SOURCE DEST ...... Verified with -132.el6: ============================================================ ...... Saving to remote location root.65.161 1+0 records in 1+0 records out Free memory/Total memory (free %): 215972 / 249524 ( 86.5536 ) Copied 0.515625 MB / 1791.38 MB ...... Change status to VERIFIED. Red Hat Enterprise Linux 6.0 is now available and should resolve the problem described in this bug report. This report is therefore being closed with a resolution of CURRENTRELEASE. You may reopen this bug report if the solution does not work for you. |