Camelia, the Perl 6 bug

IRC log for #gluster-dev, 2013-04-19

| Channels | #gluster-dev index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:18 awheeler_ joined #gluster-dev
00:19 hagarth joined #gluster-dev
00:23 sghosh joined #gluster-dev
00:37 hagarth joined #gluster-dev
00:59 hagarth joined #gluster-dev
01:01 yinyin joined #gluster-dev
01:29 itisravi joined #gluster-dev
02:00 yinyin joined #gluster-dev
03:02 vshankar joined #gluster-dev
03:29 awheeler_ kkeithley: getting the xattr error
03:35 awheeler_ kkeithley: commented out the requirement, and now things are kind of working -- GETS don't work, but PUTS and DELETES are fine.
03:36 awheeler_ Apr 19 04:33:21 r3c1node0 proxy_server ERROR with Object server 127.0.0.1:6010/5586aea39df84582887901c2d7488db0 re: Trying to GET /v1/AUTH_f645da458e454593a72d7b6d6856​b3b3/685c0fc275964b69bc6dce33ede2232d​_3/711d83ae6b554c318cd2afb5fceaa31e: #012Traceback (most recent call last):#012  File "/usr/lib/python2.6/site-packages/​swift/proxy/controllers/base.py", line 795, in GETorHEAD_base#012    possible_source = conn.getresponse()#012  File "/usr/lib/
03:36 awheeler_ python2.6/site-packages/swi​ft/common/bufferedhttp.py", line 102, in getresponse#012    response = HTTPConnection.getresponse(self)#012  File "/usr/lib64/python2.6/httplib.py", line 990, in getresponse#012    response.begin()#012  File "/usr/lib64/python2.6/httplib.py", line 391, in begin#012    version, status, reason = self._read_status()#012  File "/usr/lib64/python2.6/httplib.py", line 355, in _read_status#012    raise BadStatusLine(line)#
03:36 awheeler_ 012BadStatusLine (txn: txdc7723d6bb534c928dd405afb8fcac37) (client_ip: 172.16.1.93)
03:49 portante joined #gluster-dev
03:55 itisravi joined #gluster-dev
03:58 itisravi joined #gluster-dev
04:18 itisravi joined #gluster-dev
04:24 itisravi joined #gluster-dev
04:24 sgowda joined #gluster-dev
04:26 itisravi joined #gluster-dev
04:44 hagarth joined #gluster-dev
04:46 aravindavk joined #gluster-dev
04:50 raghu joined #gluster-dev
05:14 bala joined #gluster-dev
05:25 mohankumar joined #gluster-dev
05:25 deepakcs joined #gluster-dev
05:31 lalatenduM joined #gluster-dev
05:52 pai joined #gluster-dev
06:19 rgustafs joined #gluster-dev
06:39 rastar joined #gluster-dev
06:44 johnmark joined #gluster-dev
07:13 puebele joined #gluster-dev
07:41 aravindavk joined #gluster-dev
07:46 mohankumar joined #gluster-dev
07:54 bulde joined #gluster-dev
08:44 spai joined #gluster-dev
08:47 edward1 joined #gluster-dev
08:52 bulde1 joined #gluster-dev
09:56 itisravi joined #gluster-dev
11:34 bulde joined #gluster-dev
12:15 hagarth joined #gluster-dev
12:19 spai joined #gluster-dev
12:35 awheeler_ joined #gluster-dev
12:38 awheeler_ kkeithley: ping
12:39 kkeithley hi
12:40 awheeler_ Morning. :)
12:40 awheeler_ Tried out the 3.4 alpha 3.02
12:40 kkeithley yes
12:40 kkeithley and
12:40 kkeithley you got the xattr error
12:40 awheeler_ yes, and GETS don't work, but PUTS and DELETES are fine
12:41 awheeler_ getting ready to run the unit tests to see if they fail somewhere
12:41 kkeithley I used to think that I was going to like python
12:42 awheeler_ lol, I imagine you would if you could stick with Fedora.  :)
12:42 awheeler_ I see the tox stuff isn't in there -- is that not approved sufficiently yet?
12:43 awheeler_ oh, hmm, it's under swift dir
12:46 awheeler_ Getting 1 error and 2 failures
12:47 awheeler_ https://gist.github.com/awheeler/5420109
12:47 awheeler_ Hmm, the ring tests are not there
12:48 awheeler_ So, those tests seem to be missing
12:49 awheeler_ That suggests that the ring patches aren't in this RPM
12:50 kkeithley they're not in any of the 3.4.0alpha3 rpms then
12:50 awheeler_ yup
12:51 kkeithley but it works on fedora
12:51 awheeler_ hmm, did you run the unittests?
12:51 awheeler_ those patches re-introduce multi-volume support
12:52 kkeithley not as part of spinning the rpms
12:53 awheeler_ ah, so works probably translates to supports just one volume.
12:53 awheeler_ which did work just fine.
12:53 kkeithley yeah
12:53 awheeler_ So, the issues with rhel 6 are probably not related directly to that.
12:53 kkeithley I'm reasonably certain we did not get the multi-volume fix in for alpha3
12:54 awheeler_ I had thought it was approved -- what remains to be done to close that ticket/review?
12:54 awheeler_ Been meaning to ask that.
12:54 kkeithley that's one of the warts with this whole gerrit, jenkins work flows. Approved != merged
12:54 awheeler_ oh, hmmm.
12:55 awheeler_ so, what about 3.3.1-13?
12:55 awheeler_ not there either.  :(
12:55 kkeithley -12 and -13 only added a fix for rdma
12:56 jclift joined #gluster-dev
12:57 kkeithley juggling community, RHS, our two python/UFO gurus are at OpenStack Summit this week, yada, yada, yada
12:57 awheeler_ bummer, and I was all excited about my patches being in someone else's RPMs.  lol
12:57 kkeithley they'll get there, trust me. ;-)
12:58 awheeler_ ok, so, hmmm.  I thought 3.4 alpha2 had part of the ring patches though.
12:58 kkeithley or your fix might be on master branch and needs to get cherry-picked to the 3.4 and 3.3 branches
12:58 awheeler_ It is in master.
12:58 kkeithley If it was in alpha2 it should still be in alpha3
12:58 awheeler_ So probably wasn't in alpha2, but was in master.
12:58 kkeithley oh, okay. So we probably need to cherry-pick it to 3.4
12:59 awheeler_ Yup, and 3.3.1-14?  :-p
13:01 kkeithley on folsom?
13:02 awheeler_ yes, the fix is swift agnostic
13:03 awheeler_ Originally built against 3.3.1-11
13:04 kkeithley that's good. that certainly improves the odds. Is it a big patch? (I _try_ not to have big patches in the fedora rpms.)
13:05 awheeler_ When would it stop being a patch?  It's in main, so seem like it would eventually, no?
13:05 kkeithley yeah, if it's in main and we merge/cherry-pick it to the release-3.3 branch then it'll be in 3.3.2. Likewise for 3.4
13:06 kkeithley It'll be in 3.3.2 without an rpm build patch
13:06 awheeler_ cool
13:06 jdarcy joined #gluster-dev
13:07 awheeler_ Looks like it never did get a +2
13:07 awheeler_ But it does say merged
13:09 awheeler_ I guess 8k is a decent sized patch
13:10 kkeithley Let me try to solve this alpha3 on rhel issue, then I can look at a fedora/epel 3.3.1-14 with your ring patch, then do a cherry-pick to release-3.3 and release-3.4 branches so that the fix will be in the next releases.
13:10 kkeithley size is subjective
13:10 awheeler_ sounds like a plan.  :)
13:17 jdarcy Which patches are these?
13:17 kkeithley awheeler_: do you have the gerrit change ID handy?
13:28 jclift kkeithley: Did that mellanox adapter arrive?
13:29 kkeithley it arrived yesterday
13:29 jclift Cool.
13:33 jclift awheeler_: Is this the patch you guys are talking about? http://review.gluster.org/#/c/4781/
13:34 awheeler_ jclift: No, the multi-volume patch - http://review.gluster.org/#/c/4748/
13:34 mohankumar joined #gluster-dev
13:34 jclift awheeler_: k
13:35 jclift awheeler_: Have you taken a look at the glupy stuff yet?
13:36 jclift awheeler_: i.e. http://www.linuxjournal.com/con​tent/extending-glusterfs-python and https://github.com/jdarcy/glupy
13:37 * jclift suspects it might be your kind of thing too :D
13:37 awheeler_ nope, forgot about it.  :)  have to look again
13:38 jdarcy We have a regular intern plus a GSOC intern working on it in BLR.
13:38 * jclift wants to try doing stuff with it too soon
13:38 jdarcy I'll keep an eye out for those pull requests.  ;)
13:38 jclift :)
13:38 jdarcy Actually I should be moving that to the Gluster forge.
13:39 jclift jdarcy: johnmark mentioned the other day that Gitorious can mirror projects in GitHub
13:39 jclift jdarcy: So, that might be an easy way to do it?
13:40 jdarcy Maybe.  Haven't looked into it yet.
13:44 kkeithley puzzle #1. 3.4.0-2.alpha3 on real rhel6, but with unpatched pip-requires (i.e. .../swift-1.8.0-py2.6.egg-info/requires.txt with xattr>=4)  falls downpkg_resources.DistributionNotFound: xattr>=0.4
13:44 jclift jdarcy: No worries.  Just mentioned it so you know for later, when you get around to actually needing to have it all done. :)
13:44 jclift Oops, airport time.
13:45 kkeithley reverting to the .../swift-1.8.0-py2.6.egg-info/requires.txt with pyxattr>=0.4 (as originally patched in the 3.4.0-0.2alpha3) starts up. On to the next thing
13:45 kkeithley s/xattr>=4/xattr>=0.4/
13:47 hagarth joined #gluster-dev
14:12 wushudoin joined #gluster-dev
14:14 jdarcy Is it just because I'm in the office today, or is review.gluster.org utter $#@! today?
14:14 kkeithley awheeler_: not sure what's up. 3.4.0-0.2.alpha3 rpms should have patched .../swift-1.8.0-py2.6.egg-info/requires.txt from xattr>=0.4 to pyxattr>=0.4. Not sure why it didn't.
14:15 awheeler_ ok, weird.  Also, when building from source, I get an error about the filter option.
14:15 kkeithley Once I change it — manually atm — then swift starts and I can create containers and upload and download files
14:16 awheeler_ so you aren't seeing the bug I was seeing
14:16 awheeler_ ?
14:16 kkeithley apparetnly not
14:17 awheeler_ hmm
14:17 awheeler_ good, just me then.  :)
14:23 kkeithley don't you just hate that? ;-)
14:24 awheeler_ Did you try multiple volumes?
14:24 awheeler_ No, because that patch isn't there.  That could be it.
14:24 kkeithley er, no, (because your mv fix isn't in)
14:26 awheeler_ yup.  Do you use swift bench for you testing?
14:26 kkeithley I'm just doing a few curl cmds by hand
14:27 awheeler_ 3 unit tests have issues, so strange you don't see an issue.
14:27 awheeler_ but perhaps it's just me -- I didn't start with a clean build.
14:43 spai joined #gluster-dev
14:53 spai joined #gluster-dev
15:05 kkeithley awheeler_: well #else != %else in a spec file. That's why the 3.4.0-0.2alpha3 that I pointed you at had the xattr breakage. sigh
15:07 jbrooks joined #gluster-dev
16:00 awheeler_ ah, that would do it.
16:01 awheeler_ kkeithley: If you look at the last comment on the review, it says the change had been successfully cherry picked: http://review.gluster.org/#/c/4748/
16:01 awheeler_ Is that just a reference to be included in master?
16:17 ndevos awheeler_: you can do 'git branch -r --contains $githash' if your repo is up to date
16:18 awheeler_ ndevos: what will that do?
16:19 ndevos I tend to 'git fetch -v --all' to get all branches
16:19 ndevos awheeler_: it will print the remote branch that contains the $githash - that $githash is the cherry-pick hash in the last review comment
16:20 awheeler_ ah, ok, so it came back with origin/HEAD -> origin/master, and origin/master
16:20 ndevos that looks like it's in master then :)
16:20 awheeler_ any way I can use that hash to create a diff to apply to another branch, say 3.3 or 3.4?
16:21 ndevos when 3.5 gets branched off, you will see release-3.5 as well
16:21 awheeler_ Or likely the previous dependent patches have their own hashes.
16:21 ndevos uh, huh?
16:22 awheeler_ It's in master, but not in the 3.3 or 3.4 branches
16:22 ndevos yes, if you need it in 3.3 and 3.4 as well, you will need to file review requests for those branches too
16:25 ndevos which is like: git checkout -t -b my-feature-3.4 release-3.4 ; git cherry-pick $githash ; git commit --amend ; ./rfc.sh
16:26 * ndevos signs out for the day, good luck!
16:35 hagarth joined #gluster-dev
16:37 kkeithley awheeler_: IIRC, that cherry-pick means that the change was cherry-picked from the gerrit review branch to master/HEAD. As ndevos indicates, we need to file review requests to merge/cherry-pick the change to the release-3.[34] branches
16:41 awheeler_ right, cool
16:45 awheeler_ oh, cherry-pick is a command, awesome.
17:00 awheeler_ ok. so I think I have the 3.4 merge.  need to check and then submit
17:02 portante joined #gluster-dev
17:18 awheeler_ How do I cherry-pick ins the original patch which wasn't from me?
17:20 awheeler_ For the original author
17:22 awheeler_ kkeithley: ^^^ jdarcy has asked me to do the backport, so I'm taking a stab at it.
17:56 awheeler_ Ok, I have successfully created http://review.gluster.org/4857, and http://review.gluster.org/4858
17:59 awheeler_ kkeithley: ^^^
18:01 kkeithley okay
18:03 portante joined #gluster-dev
18:04 awheeler_ Once you tell me they look generally correct, I'll do the same for 3.3
18:05 awheeler_ Not sure how to indicate original author, and the bugid referenced is the RHS one.
18:06 kkeithley yes, I think that's why it failed the rh-bugid test
18:06 awheeler_ not sure how to fix that
18:06 kkeithley maybe the easiest way is to open a new BZ
18:07 kkeithley against glusterfs
18:07 kkeithley BZ 924792 is the GlusterFS BZ
18:08 awheeler_ I did, which is bz that blocked the original, but the hash is against the original.
18:08 awheeler_ hmm
18:08 awheeler_ It's the content of the hash commit that caused the issue, I think.
18:08 awheeler_ there is no hash for the non RH version, I think.
18:09 awheeler_ well, at least I didn't see it in the review
18:09 kkeithley you're referring to the git commit hash, yes?
18:09 awheeler_ yes, for the cherry picking
18:09 awheeler_ so, best thing is to find the new hash and abort this review?
18:09 kkeithley well, if that hash is in the glusterfs git repo....
18:10 awheeler_ it is, just looked: https://github.com/gluster/glusterfs/commi​t/598ca6bbaabc0b67708a1ecfbef1372eb9927ed9
18:10 awheeler_ with the RHS bug id
18:10 kkeithley maybe abort this review. save your original tree and copy the changed files into the new tree.
18:10 kkeithley yeah, this is git black magic AFAICT
18:10 awheeler_ do, don't do the cherry pick?
18:11 awheeler_ This is what you were going to do, though, yeah?
18:11 kkeithley check out a fresh tree, copy the changed files from your first tree
18:11 kkeithley yes
18:11 kkeithley wait, which?
18:11 awheeler_ What about authorship?  That history is lost.
18:12 awheeler_ You mentioned you would backport the changes to 3.3 and 3.4 after fixing the rhel 6 xattr bug.
18:12 kkeithley I'm not so worried about authorship
18:12 awheeler_ Then jdarcy asked me if I needed help doing it.
18:12 kkeithley yes. but if cherry-pick isn't going to cooperate...
18:13 kkeithley you can always credit Junaid in the commit message
18:13 awheeler_ Ok. then I'll do as you suggest -- and I'll cherry pick my changes.   I guess that means two commit still, and two reviews?
18:14 awheeler_ Can we just ignore the bugid failure?
18:15 kkeithley gerrit or jenkins updates the BZ with the commit hash. This is there to prevent updating the wrong BZ, especially when the wrong BZ isn't even GlusterFS or RHS, so....
18:16 awheeler_ Ah, hmm.  Could I amend the commit message and change the bugid?
18:16 kkeithley that might work, try it and see.
18:18 awheeler_ I still have to abandon though.
18:18 kkeithley you can definitely amend the commit and change the BUG. Then see if it passes the rh-bugid test
18:18 kkeithley why?
18:19 awheeler_ Oh, (too late) because the command that needs amending is not the latest commit
18:19 awheeler_ can I un abandon?
18:19 awheeler_ looks like I can.
18:19 awheeler_ So, how do I amend the commit if it's not the most recent one?
18:20 kkeithley git commit -a --amend isn't working?
18:21 kkeithley This is more git black magic. Someday I'll understand
18:21 awheeler_ Nope, just the latest commit shows up
18:21 awheeler_ And since I cherry-picked, I couldn't do them together.
18:22 awheeler_ not a problem to re-create it.
18:22 awheeler_ (I've already gone through this process for this change about 10 times, starting to get it. :)
18:23 kkeithley What I've done before is: git clone .../glusterfs.git; cd glusterfs.git; <git fetch or git pull> $changeset; git cherry-pick $hash
18:24 kkeithley Looking at it I realize the gerrit review page has a cherry-pick line that will combine the fetch and cherry-pick in one.
18:24 kkeithley s/line/example/
18:25 awheeler_ not quite following you.  Where is the release-3.4 checkout?
18:25 awheeler_ Or would you be starting from where I am now?
18:26 awheeler_ and what is changeset in this case?  Not the same as the cherry pick yes, right?
18:27 kkeithley just closed the window.  hang on
18:28 kkeithley fwiw, using cherry-pick is supposed to be a convenience in git. It's not mandatory that we use it for this
18:28 kkeithley sorry, yes, I did not show the release-3.4 checkout
18:29 kkeithley git clone, cd glusterfs, git checkout release-3.4, git fetch...
18:30 awheeler_ but I can't use the existing reviews, right?
18:30 kkeithley no, we need new ones, but they should be pro forma at this point
18:30 kkeithley or yes, we need new ones, but they should be pro forma at this point
18:31 awheeler_ lol, got to love the ambiguity of English sometimes.
18:37 awheeler_ strange the changeid is not in the right place.
18:40 awheeler_ having trouble getting rfc.sh to accept it.  Perhaps the bugid change isn't allowed?
18:40 awheeler_ error: failed to push some refs to 'ssh://awheeler@git.gluster.org/glusterfs.git'
18:41 awheeler_ or is that because I already have tried a push for that branch?
18:41 kkeithley Dunno, maybe wave a dead chicken over it?
18:43 awheeler_ It was trying to push to the same reviews, and they were closed.
18:43 awheeler_ Re-opened, and now it says I
18:43 awheeler_ 'm not making any changes.  lol
18:43 awheeler_ so, now what?
18:43 kkeithley I'm trying, let's see what happens
18:44 kkeithley it's just bits and cpu cycles
18:44 * kkeithley doesn't care how they get in as long as they get in
18:45 awheeler_ lol, indeed.  I could edit the commit message on the review server, to fix the bugid.  But I'll wait to see how it goes for you.
18:46 kkeithley original change was http://review.gluster.org/#/c/4485/, right?
18:46 kkeithley looks like it
18:47 awheeler_ yup, looks like it
18:48 awheeler_ You want to handle the 3.3 version?  Or shall I start work on that one?
18:51 kkeithley bazinga
18:52 kkeithley up to you
18:53 kkeithley wait, where'd it go?
18:53 awheeler_ ?
18:54 kkeithley oh, okay, it's under your change set. http://review.gluster.org/#/c/4857/
18:54 kkeithley let's see if it passes the rh-bugid test
18:55 kkeithley see if it looks right
18:56 awheeler_ yup.  The dependent review is showing outdated now.
18:58 kkeithley what do we need to do there? Is that the other one?
18:59 kkeithley you probably don't know any more than I do
18:59 awheeler_ It passed! Thanks.
18:59 kkeithley review that puppy
18:59 awheeler_ I'm hoping it will just treat that as warning info.
19:00 awheeler_ Done.
19:00 kkeithley I'll just do the 3.3 one as well then?
19:00 awheeler_ Sounds good to me.  :)
19:01 awheeler_ It's a little more involved anyway.  ;-p
19:01 kkeithley oh, sure. ;-)
19:02 awheeler_ There are more dependent hashes you would need to cherry-pick, I think.
19:07 kkeithley can I just take the changed files from the release-3.4 branch? Or make diffs and apply them on the release-3.3 branch?
19:08 kkeithley earlier you said your changes were swift version agnostic
19:10 kkeithley never mind, the same cherry pick that worked on release-3.4 has conflicts on release-3.3
19:10 kkeithley might need to rethink this
19:12 awheeler_ I can fix that i think.
19:13 awheeler_ not sure why it's saying to .gitignore ufo though
19:14 awheeler_ here goes ...
19:16 kkeithley after all was said and done, I did a git commit -s -a --amend --author=$me, followed by the ./rfc.sh
19:16 awheeler_ for 3.4?
19:16 kkeithley yes
19:16 kkeithley yes for 3.4
19:16 awheeler_ after you updated the bugid in the commit?
19:17 kkeithley right
19:19 awheeler_ strange about the changeid.  I keep having to paste it at the bottom'
19:19 kkeithley I didn't do anything with the changeid, I think you're going to get a new one
19:20 awheeler_ perhaps, but since I don't have one during the commit, I'm just copying the existing changeid already in the commit.
19:20 awheeler_ which is probably wrong
19:21 awheeler_ so, should I append the changeid that it's in the sample output from the failed import?
19:21 awheeler_ s/import/rfc.sh/
19:22 awheeler_ seems to work
19:22 awheeler_ and we have two new reviews: http://review.gluster.org/4860 and http://review.gluster.org/4861
19:23 awheeler_ now we just have to see what the build thinks of it.
19:23 awheeler_ Would be nice to get junaid to bless it as he's the original author.
19:25 ddia joined #gluster-dev
19:28 noche hi, when I type /usr/sbin/gluster volume rebalance gv0  status --xml I get plain text and xml on the same response (http://pastebin.com/wqW9XA6Q)
19:29 noche it is version 3.3.1 on centos
19:29 awheeler_ I can confirm that I see the same thing.
19:40 awheeler_ looks like it might be fixed in 3.4, lots of code changes in that area
20:03 kkeithley awheeler_: your release-3.3 change passed all the build tests. and we can add junaid to the reviewers.
20:04 awheeler_ I see that, and I did add Junaid, so cool.  Looks like the 2nd part of 3.4 won't get tested until it's somehow shown to be not out-of-date
20:08 kkeithley Did you add Junaid to both or just your release-3.3?
20:08 kkeithley let me look
20:08 awheeler_ fixed
20:08 awheeler_ just did a git commit --amend and then a new rfc.sh
20:09 awheeler_ adding him to 3.4
20:10 kkeithley cool. Well, it's almost Miller Time. Let's see if I can get home or whether there'll be a police barricade at the freeway offramp
20:11 awheeler_ good luck!
20:12 awheeler_ BTW, I've updated my openstack-keystone glusterfs doc to use the version from RDO, and no longer need the race condition patch.
20:16 kkeithley Yes, I need to get to that too
20:19 awheeler_ Would there be value in creating an AWS image pre-installed with that?
20:21 kkeithley oh definitely
20:23 awheeler_ Ok, then I'll look at doing that and adding it to my doc.
20:35 awheeler_ Damn, looks like when I updated 3.3, somehow 3.4 got submitted too.
20:37 awheeler_ re-fixed the commit message.
20:44 awheeler_ Back in business.  :)
20:50 awheeler_ now back out of business.
22:05 xavih joined #gluster-dev
22:11 awheeler_ got accepted again, but requires new review.
22:25 johnmark awheeler_: thanks for updating the doc
22:25 johnmark you are special!
23:50 awheeler_ joined #gluster-dev
23:51 awheeler_ johnmark: This time I followed it to build a working system, and fixed the issues there, so it should be in pretty good shape.

| Channels | #gluster-dev index | Today | | Search | Google Search | Plain-Text | summary