-
Notifications
You must be signed in to change notification settings - Fork 77
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
noobaa log is not performing logrotate #8153
Comments
@rkomandu please check the logrotation when it hits the size limit in your node, Last time when you did the upgrade log size was already larger than configured value and I believe logrotate failed because of that. If logrotate successful when |
Ran IO workload onto the app2 node serving the IP, logrotate didn't happen
|
rsyslog on the node
|
@rkomandu Could you please share the output for command |
ls -l /etc/logrotate.d/noobaa/logrotate_noobaa.conf |
@naveenpaul1 , as we discussed in our rsyslog call regarding the log rotate not functioning as expected. Below is on a fresh BM system, the log hasn't been rotated automatically
|
@naveenpaul1 Hence there are two things here (which can be interrelated w/r/t logrotate not happening, hence the noobaa.log can't add more data to the current size of 106M). Request you to look through this as a priority to get workaround for our GA (i know we can run manually logrotate command), but still we might loose some data once log reaches that size limit. @guymguym this has to have solution for us and fix this as part of the overall logging implementation later on. |
@naveenpaul1 , as per our slack discussion (https://ibm-systems-storage.slack.com/archives/D06A78LUJUQ/p1719223293820039?thread_ts=1719219424.304859&cid=D06A78LUJUQ) noobaa-5.17 rpm is installed and still we see issue w/r/t some minutes are kind of lost while executing the IO benchmark (rpm: noobaa-core-5.17.0-20240624.el9.x86_64 --> noobaa-core-5.17.0-rsyslog-fail-bm-20240624.el9.x86_64.rpm ) It is stopping after every few mins (in noobaa.log --> we see this as follows first time after 02:17:14 --> 02:26:00 no update to noobaa.log
Second time no update to the noobaa.log (it doesn't update, same is showed above for about 8m 41sec i.e 02:27:20 --> 02:36:01)
timestamps as well (02:31:22 EDT and 02:34:17 EDT)
As you can observe from above noobaa.log, there is no log entry after 02:27:20 --> 02:36:01 (all of them are lost) , it is happening repeatedly . Fix has to be checked further why does it halt in between for updating the logs while IO is still running, as the logging should be continuous
|
@naveenpaul1 and I had been on the BM and saw the system didn't get through with the rotation, so Naveen will analyze further and come back on the fix. this logging is an important fix required from our customer / in-house issues for tracking, loosing them in between as posted in my last comment is not a good sign as well. |
I was checking the issue in BM shared by @rkomandu.
and looks like script( |
@rkomandu @naveenpaul1 Fixed by Naveen on #8182, closing and adding request validation label |
@romayalon @guymguym |
we should delete the /etc/logrotate.d/noobaa directory
Aren't we having a single file ? it is confusing for the customer as well |
@rkomandu |
@naveenpaul1 I don't see the same on the app3 and app4. please ensure to clean it up Till now once the noobaa rpm is removed/erased, the respective files are cleared all the time. |
@rkomandu That file was created manually for testing that's why it didnt got removed by rpm uninstall, You can ignore it for now. |
@naveenpaul1 , so there is no limit of 105M now looks like with the new rpm
|
@rkomandu When logrotate rotate the noobaa.log at midnight it will use |
logrotate did take place at 23:59
As discussed with you at length over slack, the app3, app4 nodes didn't have that extra file (logrotate_noobaa.conf) which was manually created by you on app2. rpm -qf /etc/logrotate.d/logrotate_noobaa.conf rpm: noobaa-core-5.17.0-20240704.el9.x86_64 |
Thank you @rkomandu for your confirmation. |
Environment info
Noobaa staging build of 4.15.4 (noobaa-core-5.15.4-20240611.el9.x86_64)
Actual behavior
When the noobaa.log reached the 105M, it should automatically rotate as per the logrotate setting, it doesn't seem to be the case on the BM (physical machine). @naveenpaul1 and I looked at the system in detail and after running manually (restarting rsyslog) it worked.
All the discussion is in this thread https://ibmandredhatguests.slack.com/archives/C015Z7SDWQ0/p1718358615847589 (Guy and Naveen)
after performing this
Where we haven't performed the rsyslog restart, here it is
Expected behavior
It should automatically logrotate when the file size is reached.
Steps to reproduce
More information - Screenshots / Logs / Other output
Posted above the log file
The text was updated successfully, but these errors were encountered: