| Summary: | sstate create package issue | ||||||
|---|---|---|---|---|---|---|---|
| Product: | [Build System, Metadata & Runtime] Meta-yocto | Reporter: | Matthew McClintock <msm-oss> | ||||
| Component: | meta-yocto | Assignee: | Richard Purdie <richard.purdie> | ||||
| Status: | RESOLVED FIXED | QA Contact: | |||||
| Severity: | normal | ||||||
| Priority: | Low | CC: | aprasad, msm-oss, poky.bs.watcher, poky.watcher | ||||
| Version: | 1.1 | ||||||
| Target Milestone: | --- | ||||||
| Hardware: | x86 | ||||||
| OS: | Multiple | ||||||
| Whiteboard: | |||||||
| OS type for building Yocto: | --- | Type of Regression: | --- | ||||
| Verified: | Documentation change: | --- | |||||
| Attachments: |
|
||||||
|
Description
Matthew McClintock
2011-11-19 11:33:18 UTC
Do you have any more logs from when this happened. It sounds like the gzip binary in the sysroot is being updated (hence it being busy) at the same time as something else was trying to run it. Created attachment 282 [details]
full console log of failure
(In reply to comment #1) > Do you have any more logs from when this happened. It sounds like the gzip > binary in the sysroot is being updated (hence it being busy) at the same time > as something else was trying to run it. Anyone else besides this? The log in tmp/work/ where just the same thing thats in this console message No, that is perfect, thanks. You'll notice the line: 14: gzip-native-1.4-r0 do_populate_sysroot (pid 20711) which confirms my suspicion about what this was racing with. The trouble is we don't list gzip-native in dependencies and if we did, it would be everywhere although I notice its not in ASSUME_PROVIDED. Perhaps the correct solution is to put it there and also not to install the gzip binary in gzip-native. Looking at the code: # don't use version earlier than 1.4 for gzip-native, as it's necessary for # some packages using an archive format incompatible with earlier gzip PREFERRED_VERSION_gzip-native ?= "1.4" and expat does: do_unpack[depends] += "gzip-native:do_populate_sysroot" undoubtedly due to file format issues we've had in the past. Not an easy problem... http://git.yoctoproject.org/cgit.cgi/poky/commit/?id=0ff62b0462f3f64672bd4704de9a192eb1a730d1 The solution might be to treat this the same way as we treat tar-native (check version and build right at the start if needed). *** Bug 1822 has been marked as a duplicate of this bug. *** I think I have a way to resolve this kind of problem. We should do something similar to http://git.yoctoproject.org/cgit.cgi/poky-contrib/commit/?h=rpurdie/t1&id=9cdb7194592c1bec5d332268d1be3e736dbcdd47 for the package creation step, removing binaries in the sysroot from the equation. This should be fixed by: http://git.yoctoproject.org/cgit.cgi/poky/commit/?id=40f95685e6a3b4359d2ab2fea1418fe5e4bd0aa1 |