Nir Soffer has posted comments on this change.

Change subject: blockSD: Fix busy loop when zeroing image volumes
......................................................................


Patch Set 1:

I think this may be the the root cause for network disconnects when copying 
disks on block storage - once the copy finish we wipe the old disks, mixing 
crazy amount of io (multiple dd processes pushing zeros to storage), and doing 
millions of syscalls, looping over poll(), waitpid() epoll() in AsyncProc etc.

Needs more testing, but I have seen too many disconnect lately when testing 
disk copies,and vdsm is always running with very high cpu usage during this 
time.

-- 
To view, visit https://gerrit.ovirt.org/57541
To unsubscribe, visit https://gerrit.ovirt.org/settings

Gerrit-MessageType: comment
Gerrit-Change-Id: I7f5ec5faefb6840f1bd5348b93cc784d0aff8690
Gerrit-PatchSet: 1
Gerrit-Project: vdsm
Gerrit-Branch: master
Gerrit-Owner: Nir Soffer <[email protected]>
Gerrit-Reviewer: Adam Litke <[email protected]>
Gerrit-Reviewer: Allon Mureinik <[email protected]>
Gerrit-Reviewer: Dan Kenigsberg <[email protected]>
Gerrit-Reviewer: Freddy Rolland <[email protected]>
Gerrit-Reviewer: Idan Shaby <[email protected]>
Gerrit-Reviewer: Jenkins CI
Gerrit-Reviewer: Nir Soffer <[email protected]>
Gerrit-Reviewer: Piotr Kliczewski <[email protected]>
Gerrit-Reviewer: Yaniv Bronhaim <[email protected]>
Gerrit-Reviewer: gerrit-hooks <[email protected]>
Gerrit-HasComments: No
_______________________________________________
vdsm-patches mailing list
[email protected]
https://lists.fedorahosted.org/mailman/listinfo/vdsm-patches

Reply via email to