SymptomsAfter a server crash, one may encounter the "Blue Screen of Death" (BSOD) with a Windows server, or a kernel panic with a Linux server. BSODs are accompanied with System Event log messages like the following:
The previous system shutdown at time on date was unexpected.
Kernel panics can be identified by kernel messages on the server console:
Code: 31 c0 89 14 24 90 0f ab 02 b9 10 00 00 00 89 d8 e8 a4 b4 0c 00 89 d8 83 c4 1c 5b 5f c3 53 b8 00 e0 ff ff 21 e0 8b 48 10 83
c8 ff <87> 05 e0 d0 ff ff 8b 15 d0 d0 ff ff bb d0 d0 ff ff 0f b6 81 a0
<0>Kernel panic - not syncing: Attempted to kill init!
<1>Unable to handle kernel paging request at virtual address ffffd370
Crashes can be divided into two categories:
- Periodical - When a crash happens repeatedly or there are definite steps to reproduce the issue
- Rare - When the crash happened once and there is no stable sequence of actions to reproduce the crash
Collect necessary information
If the server crashes rarelyIf the crash cannot be reproduced or crashes no longer occur, then it is reasonable to configure memory dump generation:
- For Windows, the process is described in KB article 10055.
- For Linux, it is necessary to configure Kdump as per KB article 10044.
If the server crashes frequently or there are steps to reproduce the issueIf the server crashes very often or there is a definite sequence of steps to trigger the crash, then it is necessary to configure a serial console connection first:
After the above configuration is applied and proved to be working, it is sufficient to provide Parallels Support with access to the server suffering from hangs, access to the Monitoring Node, and with the sequence of actions to reproduce the hang.
Additional information needed by support
Answers to these questions:
- How often is the crash reproduced?
- Was any new program installed before the crash first occurred?
- Were any software updates installed? OS updates or patches?
- Were any devices added to the server? New drivers?
- Which steps or actions preceded the crash? Is it possible to use these steps to reproduce the crash again?
All relevant logs can be submitted via built-in tools:
On Parallels Containers server:
1270 How do I submit a Support Issue?
On Parallels Server Bare Metal:
# prlsrvctl problem-report --send
- On Parallels Containers server:
- Compressed memory dumps
- Access to Hardware Node
- Access to Monitoring Node
Submit information to support
- If you do not have a support ticket open, please create one first.
- If you have the ticket, provide an issue description in that ticket.
- Log files are submitted via built-in tools as described above.
- Contact a Parallels Support representative within the support ticket and request the link to upload the dump to.
Additional informationFor more information about server hang scenarios, see this KB article:
10049 How to handle a server hang scenario