Changed flags for _ltfs_search_index_wp #480
Conversation
This work will be moved forward on `bringup-25` branch for a while. And teh branch will be merged to master in the future.
- Use Debian10 on check build correctly (LinearTapeFileSystem#235) - Fix build checker break on macOS - Remove Ubuntu19(Eoan) because it is no more supported - Add a description to refresh shared library cache by ldconfig -v
- Fix of issue LinearTapeFileSystem#225 - Remove accidental committed files
Treat errno 95 on ltfs_ordered_copy to support destination file system which doesn't support xattr. Fix LinearTapeFileSystem#236, LinearTapeFileSystem#243
This change also includes following changes. - Remove warning on RHEL8 - Enhance message checker to support RHEL8 - Improve auto build checker in macOS and RHEL8 (use message checker) - Fix unknown messages
- Introduce generic logselect issue I/F via EA by ltfs.vendor.IBM.logPage.XX.YY - XX shall be a hex value like A0 as page code. YY XX shall be a hex value like 60 as subpage code - Once LTFS receives a correct VEA, XX and YY can be converted by strroul() with base=16, LTFS always issues logsense to the drive and return the specified page by binary
- Introduce generic MAM fields capture I/F via ltfs.vendor.IBM.mediaMAM.XX - XX shall be one of 00, 01 , IP or DP - The new introduced VEA returns the MAM value by binary. You need to parse it by yourself
…em#276) * Handle copy 0-byte file correctly * Fix the problem when only one file are specified as a source and `--keep-tree` is specified * Fix the destination path for copying EAs * Modify .gitignore
Co-authored-by: Atsushi Abe <piste@jp.ibm.com>
|
I would like to ask you the things below.
|
|
Could you take care about my previous comment about PR #480 ? |
|
@juliocelon , @perezle , Could you handle this? It looks issue #479 and it's fix (this PR) come from a customer. I don't know I just push this only on master branch. |
|
Nice to hear from you Abe-san.
Thanks!
Leonel Perez
Modern Data Protection, Physical Tapes & Spectrum Archive Manager
e-mail: ***@***.******@***.***>
México, Guadalajara Campus.
…________________________________
From: Atsushi Abe ***@***.***>
Sent: Monday, September 9, 2024 8:33 PM
To: LinearTapeFileSystem/ltfs ***@***.***>
Cc: LEONEL PEREZ PELAYO ***@***.***>; Mention ***@***.***>
Subject: [EXTERNAL] Re: [LinearTapeFileSystem/ltfs] Changed flags for _ltfs_search_index_wp (PR #480)
@ juliocelon , @perezle , Could you handle this? It looks issue #479 and it's fix (this PR) come from a customer. I don't know I just push this only on master branch. — Reply to this email directly, view it on GitHub, or unsubscribe. You
@juliocelon<https://github.com/juliocelon> , @perezle<https://github.com/perezle> ,
Could you handle this?
It looks issue #479<#479> and it's fix (this PR) come from a customer. I don't know I just push this only on master branch.
—
Reply to this email directly, view it on GitHub<#480 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ABX53RVHQWEH4BZRNTAASTTZVZK65AVCNFSM6AAAAABNFESV7GVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDGMZZGQ4DSOJQGI>.
You are receiving this because you were mentioned.Message ID: ***@***.***>
|
|
Hello @piste-jp, This PR should be merged to You mentioned that the file backend might be helpful to the test this kind of issue. Did you mean to use "ltfs.vendor.IBM.forceErrorWrite" extended attribute ? I was planning to use ITDT to test this change, but could you provide me more details in case I use a file backend? Sorry for the delay of this answer, I have been working in other tasks. Regards |
There was a problem hiding this comment.
@juliocelon , @perezle ,
Both of you MUST NOT merge this PR to the stable branch!!!! This PR ruins the stable branch completely.
Did you merge v2.4-stable branch to your dev branch (and push)?
It is completely wrong operation. Please fix your PR.
I believe a tape which has following condition is needed for your test.
I'm not sure how you can create a tape both condition was met. But you can use ITDT if you know how to do it obviously. My intention is the file backend manages MAM and VCR on its file, so you can create those condition with editing file backend files (while ltfs is not running). |
My branch was created from master branch so all these conflicts are because the differences between master and Regards |
Summary of changes
This pull request includes following changes or fixes.
_ltfs_search_index_wpDescription
When a tape cartridge with a write permanent error is trying to be mounted and the MAM (cartridge memory) attribute of the Index Partion (IP) stores a generation number different than the MAM attribute of the Data Partition (DP), the mount process fails even when ltfs can still search for the index in the other partition.
The code that belongs to the logic that handles WP happens on IP, defined "can_skip_ip" flag as false and it shall be true. On the other hand, the code that handles WP happens on DP, defined the "can_skip_ip" as true and it should be false. The logic is not adjusted accordingly.
Fixes #issue_no
Type of change
Checklist: