-
Notifications
You must be signed in to change notification settings - Fork 1.2k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
our latest ami coredump with c-s #380
Comments
scripts to reproduce
|
|
This is on another node of the two nodes cluster:
|
Gleb - can you pleasse take a stab at this and see if you can find why its On Mon, Sep 21, 2015 at 9:24 AM, Asias He notifications@github.com wrote:
|
Reproduced with 1 node cluster:
|
I'll create an ami without enhanced networking to see its not related On Mon, Sep 21, 2015 at 9:59 AM, Asias He notifications@github.com wrote:
|
ami : ami-63d9ab06 - same as the ami that has the issue - without enhanced On Mon, Sep 21, 2015 at 10:00 AM, Shlomi Livne shlomi@cloudius-systems.com
|
On Sun, Sep 20, 2015 at 11:52:50PM -0700, slivne wrote:
|
Not clear Gleb are you working on fixing the do_flush
On Mon, Sep 21, 2015 at 11:11 AM, Gleb Natapov notifications@github.com
|
On Mon, Sep 21, 2015 at 01:14:42AM -0700, slivne wrote:
|
Instance type ? are you using the same instance type On Mon, Sep 21, 2015 at 11:16 AM, Gleb Natapov notifications@github.com
|
On Mon, Sep 21, 2015 at 01:18:02AM -0700, slivne wrote:
|
On Mon, Sep 21, 2015 at 4:16 PM, Gleb Natapov notifications@github.com
Use c3.8xlarge for both server and c-s. If one instance of c-s can not Asias |
When you start scylla server, follow instructions here: https://github.com/cloudius-systems/scylla/wiki/Using-AWS-AMI Choose instance store 0 and store 1 as the extra two disks. |
On Mon, Sep 21, 2015 at 01:23:17AM -0700, Asias He wrote:
|
On Mon, Sep 21, 2015 at 4:28 PM, Gleb Natapov notifications@github.com
Yes, but we do not know if the disk difference maters.
Try to add another instance to stress the server.
Asias |
I started another 3 instances, Reproduced again with 2 load + 1 server:
|
btw, to allow coredump to be stored on c3.8xlagre Modify coredump.conf. Disable compress and enlarge the size limit. [fedora@ip-172-31-40-176 ~]$ cat /etc/systemd/coredump.conf Bind coredump dir to our data partition which should have 2*320 GB. sudo mkdir /data/coredump |
Can you try with this? diff --git a/transport/server.cc b/transport/server.cc
|
I'm trying
I will test yours shortly. |
@gleb-cloudius with your patch, I still see the panic. |
I am going to send a simplified version of my patch soon. Hopefully, it still will make the problem go away. |
OK On Mon, Sep 21, 2015 at 7:28 PM, Paweł Dziepak notifications@github.com
Asias |
On Mon, Sep 21, 2015 at 04:22:20AM -0700, Asias He wrote:
|
This one can be closed now. On Sun, Sep 20, 2015 at 10:53:07PM -0700, Asias He wrote:
|
The text was updated successfully, but these errors were encountered: