Open dgreuel2002 opened 1 year ago
Please provide details about your Yugabyte deployment.
I connected to database, some tables return data and some time out....
./ycqlsh
Connected to local cluster at 127.0.1.1:9042.
[ycqlsh 5.0.1 | Cassandra 3.9-SNAPSHOT | CQL spec 3.4.2 | Native protocol v4]
Use HELP for help.
ycqlsh> SELECT COUNT(*) FROM idx.piecedeal;
count
-------
4719
(1 rows)
ycqlsh> SELECT COUNT(*) FROM idx.piecemetadata ;
count
-------
3893
(1 rows)
ycqlsh> SELECT COUNT(*) FROM idx.pieceblockoffsetsize ;
OperationTimedOut: errors={'127.0.1.1:9042': 'Client request timeout. See Session.execute[_async](timeout)'}, last_host=127.0.1.1:9042
ycqlsh> SELECT COUNT(*) FROM idx.payloadtopieces;
OperationTimedOut: errors={'127.0.1.1:9042': 'Client request timeout. See Session.execute[_async](timeout)'}, last_host=127.0.1.1:9042
ycqlsh>
The Yuga GUI looks ok...shows tablets under tables
@dgreuel2002 I wouldn't trust the GUI for now. We have seen similar behaviour on a completely broken cluster. It sounds like there is something wrong with your Yugabyte node if you cannot run a Count query. I would suggest checking out https://github.com/filecoin-project/boost/discussions/1797 once and ensure that your deployment is correct. The best people to help us with this would be Yugabyte team itself. Can you please join the Yugabyte slack https://communityinviter.com/apps/yugabyte-db/register and start a new help thread there? I will also monitor the thread and keep an eye on things from Boost perspective.
@LexLuthr @dgreuel2002, any updates here, has this issue been resolved?
@davidd8 There has been no updates. I applied to be a part of Yugabyte slack and I didnt get the email. I just tried registering with another email address and it worked. I will submit the ticket on Yuga Slack now.
Can we know which ALTER TABLE statement failed? Because there are 3 in the migration.
@dgreuel2002 I am assuming you haven't started migration yet. If that is the case, then I would recommend removing the current YB deployment and doing another one from scratch. Deployment using yugabyted
is now supported so it should be an easy task.
@ddorian I am not sure how to see which statement failed? You want Yuga logs? This is the only message in the WARNING LOG. . W0220 00:38:00.328931 65125 yb_rpc.cc:335] Call yb.tserver.PgClientService.AlterTable 127.0.0.1:46512 => 127.0.1.1:9100 (request call id 378) took 60007ms (client timeout 60000ms).
@LexLuthr The migration keeps failing so i would agree that it hasn't been started yet. Just to be clear though, I went from 1.0 to 2.0 and migrated. At that point, i took deals for a solid month. Now I am trying to migrate to 2.1. Is it still safe to nuke and start from scratch if I have take some deals from boost already?!? I am a little confused here.
@dgreuel2002 Can you please get me the result of the following query using ysqlsh
?
SELECT count(*) FROM PieceTracker;
We can try to alter this table manually and let the migration move from this point.
@LexLuthr
yugabyte=# SELECT count(*) FROM PieceTracker;
count
-------
3711
(1 row)
Can you please also get me select * from goose_db_version;
?
yugabyte=# select * from goose_db_version;
id | version_id | is_applied | tstamp
----+------------+------------+---------------------------
1 | 0 | t | 2023-10-23 07:20:21.31317
(1 row)
Please run the following Postgres queries on your tables in the specified order.
ysqlsh
terminal to yugabyte@LexLuthr it times out on the first line
yugabyte=# ALTER TABLE PieceTracker DROP CONSTRAINT piecetracker_pkey;
ERROR: Timed out waiting for Alter Table
@ddorian Hey! Can you advise something from YB side here?
yugabyte=# SELECT count(*) FROM PieceTracker;
count
-------
3711
(1 row)
We are not able to alter the table at all. We are getting a timeout.
@dgreuel2002 Can you restart your YB instance and retry?
@LexLuthr I have restarted yuga and same result. See INFO log from yuga logs. INFOLOG.txt
@dgreuel2002 Can you try increasing --rpc_connection_timeout_ms=150000
of ybtserver and retrying?
Boost component
Boost Version
Describe the problem
boostd-data run yugabyte-migrate --hosts 127.0.1.1 --connect-string="postgresql://yugabyte:yugabyte@127.0.1.1:5433?sslmode=disable" --miner-address FXXXXXX Error: running postgres migrations: ERROR go migration: "20230801093422_add_piece_tracker_miner_addr.go": failed to run go migration: pq: Timed out waiting for Alter Table
I am unable to migrate my DB....help.
Logging Information
Repo Steps
I am trying to migrate from 2.0.0 to 2.1.0-rc2+mainnet+git.fa9fc84
I followed steps, when i run boostd-data it tells me i need to migrate. I run the migrate and get the timeout on alter table.