Bug 1511109
| Summary: | RFE: attempt to quiet vdo remove messages | ||
|---|---|---|---|
| Product: | Red Hat Enterprise Linux 7 | Reporter: | Corey Marthaler <cmarthal> |
| Component: | vdo | Assignee: | Thomas Jaskiewicz <tjaskiew> |
| Status: | CLOSED ERRATA | QA Contact: | Jakub Krysl <jkrysl> |
| Severity: | low | Docs Contact: | |
| Priority: | unspecified | ||
| Version: | 7.5 | CC: | awalsh, jkrysl, limershe, tjaskiew |
| Target Milestone: | rc | Keywords: | FutureFeature |
| Target Release: | --- | ||
| Hardware: | x86_64 | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Fixed In Version: | 6.1.0.67 | Doc Type: | If docs needed, set a value |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2018-04-10 15:47:46 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: | |||
|
Description
Corey Marthaler
2017-11-08 17:27:19 UTC
While I agree it's noisy, it's also a pretty rare operation, and we want to make it easy to debug if it goes wrong. So I'd like to hear from engineering and QE on their thoughts on this one before I give it a pm_ack. -louis I tend to agree with Louis. Creating and removing volumes isn't something that you're going to be doing repetitively (unless you're testing or running into issues). And if you were to run into a problem, the more info the better, IMO. Though, I will note that creating/removing an LVM volume doesn't generate so much noise, so I'm open to opinions on it for sure. I agree with Corey, so much noise might be even confusing for common user, who might think something is wrong. But also this information is crucial for debugging, so hiding it under verbosity level is probably the way to go here. vdo already has --verbose, so why not add -v /-vv / -vvv as is the custom with many other utilities? Ok, I've given a pm_ack, however we still need one from QE and devel before we should pick this one up. vdo-6.1.0.55-10 [ 5025.087355] kvdo1:dmsetup: suspending device 'vdo' [ 5025.228600] kvdo1:packerQ: compression is disabled [ 5025.233443] kvdo1:packerQ: compression is enabled [ 5025.238430] kvdo1:dmsetup: device 'vdo' suspended [ 5025.243238] kvdo1:packerQ: compression is disabled [ 5025.248101] kvdo1:packerQ: compression is enabled [ 5025.252911] kvdo1:dmsetup: stopping device 'vdo' [ 5025.260447] kvdo1:dmsetup: Setting UDS index target state to closed [ 5025.269245] kvdo1:dmsetup: Setting UDS index target state to closed [ 5025.285464] uds: kvdo1:dedupeQ: index_0: beginning save (vcn 4294967295) [ 5028.166949] kvdo1:dmsetup: device 'vdo' stopped The noisiness is gone, but there are some duplications. Is there a way to get rid of them too? Highly likely we can remove some duplications. I will take a look at it. Ok, giving back. The "Setting UDS Index target state ..." duplicate has been removed. Were there any other messages that you had in mind to remove or change? Thanks Tom. The other part is this: [ 5025.087355] kvdo1:dmsetup: suspending device 'vdo' [ 5025.228600] kvdo1:packerQ: compression is disabled [ 5025.233443] kvdo1:packerQ: compression is enabled [ 5025.238430] kvdo1:dmsetup: device 'vdo' suspended [ 5025.243238] kvdo1:packerQ: compression is disabled [ 5025.248101] kvdo1:packerQ: compression is enabled [ 5025.252911] kvdo1:dmsetup: stopping device 'vdo' As you can see the kvdo disables compression and enables it again. Twice. What is the reason for this behavior and is it possible to remove one / both of them? Both suspend and stop want to make sure we have completed all I/O requests that have been started. The pair of disabled/enabled messages are how we kick all the requests that are waiting in the packer queue and force them to complete. It is a simple change to not do this disable/enable pair if we know that there are no requests in progress. In practice, this means that the second pair will never be needed since there will be no requests in progress. Also, because the device should be idle, the first pair will also not be needed. vdo-6.1.0.72-12: [ 735.651507] kvdo0:dmsetup: suspending device 'vdo' [ 735.862196] kvdo0:dmsetup: device 'vdo' suspended [ 735.883451] kvdo0:dmsetup: stopping device 'vdo' [ 735.905596] kvdo0:dmsetup: uds: kvdo0:dedupeQ: index_0: beginning save (vcn 4294967295) [ 735.941478] Setting UDS index target state to closed [ 736.093439] kvdo0:dmsetup: device 'vdo' stopped The UDS duplicate is gone and so are all the compression messages. Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHEA-2018:0871 |