TREXNameServer.cpp(03277) : shared memory lock '/tmp/.hdb_SID_00_lock'was deleted -> stopping instance ...

[Former Member]

Hi Experts,

Since we installed HANA DB, we're facing the issue "TREXNameServer.cpp(03277) : shared memory lock '/tmp/.hdb_SID_00_lock'was deleted -> stopping instance ...".

As a result, DB server going down and we have to start once again. We found SAP note: 1984700 but it doesn't explain the root cause of this issue.

Could you please share your ideas with us?

Thanks & best regards,

Sreenu

[fabianlarav]

Hello Sreenivasa,

We had the exact same issue on our Redhat host. The root cause is on:

Why SAP HANA crashes after 10days of uptime? - Red Hat Customer Portal

Apparently it is a default cron job that cleans files on /tmp older that 10 days

Thank you for your post.

[Former Member]

Hello Fabian,

I am having the exact same issue on my Redhat host. But, I am not able to access the link you have shared.

It would be great if you could share more details on how to resolve this issue.

Thanks in advance,

Sanu

[lbreddemann]

A little googling would get you far here... linux - When does /tmp get cleared? - Server Fault

[Former Member]

Hi Lars,

Thanx for the response. I understand from the SAP Note: 1984700 that I need to remove tmpwatch from the system's cron job. But, will this result in the /tmp growing in size in the course of time?

I have now created a 'jobs.deny' file in the /etc/cron.daily and blacklisted the tmpwatch.

Kindly advice if this is the appropriate method to follow.

Thanx in advance,

Sanu

[lbreddemann]

The question is why would /tmp grow over time on your SAP HANA system?

Generally there shouldn't be any other significant processes on that machine, so what do you think the /tmp filling would be caused by?

I think it's pretty simple: SAP HANA needs that file, so make sure it doesn't get deleted.

If there are lots of other processes not cleaning up their /tmp files properly, then it might be good to check why this is happening.