-
-
Notifications
You must be signed in to change notification settings - Fork 510
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
UPDATE MVA is freezing #305
Comments
could you attach your full query to issue or put it at FTP? The write-only FTP for customer data:
Could you create index with 10 to 100 documents that matches your query and upload it? In case you do not want to upload sensitive data you might index your documents with full text fields tokens replaced by Could you also check hung daemon with GDB at level
this way
if values of m_iRowID differ that mean that daemon not hung but processes filter condition but very slow. In case your query looks like it might be better to transform it to |
The problem reproduces only in archive version of index and it weights more than 200GB. Can't reproduce it in delta or main indexes.
I have checked the index with gdb in the debug version of Manticore and I have duplicate version of the system with the index of same size and environment and such updates executes in a moment. After restart of searchd it also executes in a moment, but a bit later if hangs again.
I don't have big Do you have any suggestions what else i can check? |
Additionally in this situation searchd --stopwait not working and searchd do not accept connections with such stacktrace:
|
could you provide the exact query that cause this hung hung as attached file to this ticket or upload it to FTP? |
Uploaded main info about index and example query to ftp |
looking at query you provided I see no reason for daemon to spawn FilterOr tree might be some error that related to data. And for daemon hung on shutdown UPDATE of MVA \ JSON \ STRING attributes required exclusive lock of index that is why shutdown thread can not lock index to flush it and hung too. Could you also dump in GDB content of Could you try resent release daemon to make sure the issue still here? |
Describe the environment
Manticore Search version: Manticore 3.1.2
OS version: Centos 7
Build version:
Where is it in sources?
Describe the problem
Hello!
Sometimes i am facing the following problem - my MVA updates are infinitely freezing and after restart of searchd instance everything is fine. I have got a gdb stacktrace and i suspect, that this is connected with this thread:
What can cause this behaviour? Thanks in advance!
Steps to reproduce:
UPDATE <index_name> SET <ATR_MVA_NAME> = (1,2,3) WHERE id =
Index weights more than 200Gb and contains sensitive data, what else can i provide?
In case you've faced a crash what indextool --check returns:
Indextool says everything is fine
The text was updated successfully, but these errors were encountered: