Activity
From 08/01/2016 to 08/30/2016
08/30/2016
- 10:19 AM Bug #16: Threaded GDP Clients Fail
- I think I can reproduce this bug now. At a first look, this does not appear to be specific to the Python API, but nee...
08/25/2016
- 05:51 PM Bug #45 (Closed): Reading errors (timeout, assertion failure, end of file)
- My bad, wasn't working on the latest version. Right now, the original `gdp-reader` command with the specified argumen...
- 05:24 PM Bug #45: Reading errors (timeout, assertion failure, end of file)
- Doesn't work, still. However, this is primarily an issue of corrupt index on the log-server side, which causes it to ...
- 05:31 PM Bug #51 (Closed): Corrupt time-index causes gdplogd to crash (assertion failure)
- My bad, looks like commit:b817c441 fixes this already. Should've checked out the latest version
- 05:24 PM Bug #51 (Closed): Corrupt time-index causes gdplogd to crash (assertion failure)
- How to reproduce:
```
./gdp-reader -D *=10 -t -f 1551471 -n 1 -G gdp-01.eecs.berkeley.edu edu.berkeley.eecs.swarm...
08/24/2016
- 02:56 PM Bug #50 (Closed): router will sometimes connect to itself
- I found a case on gdp-04 where the router had opened a connection to itself. The routing loop was detected but not r...
- 02:46 PM Feature #46: Need way of rebuilding log indices ("fsck for logs")
- We have some known corrupt logs right now, so this really needs to get fixed.
- 10:44 AM Feature #46: Need way of rebuilding log indices ("fsck for logs")
- See Issue #47 for a brief discussion of the issues.
- 02:45 PM Feature #43 (Rejected): A default gdp router should be contacted if a local gdp router cannot be ...
- Compiling local hostnames into the code is a bad idea. The installation scripts should populate the `swarm.gdp.route...
- 02:32 PM Bug #45 (Feedback): Reading errors (timeout, assertion failure, end of file)
- This is actually several problems. I think I have them all fixed.
* The gdp library was releasing a GCL handle wh... - 10:42 AM Bug #47 (Resolved): No sync to disk after timestamp index update
- `DB->sync` does an `fsync` system call, which is slow and expensive. Doing three on every write (since it would have...
- 10:29 AM Bug #49 (Resolved): Darwin: shared libraries have the wrong internal name
- I have changed the libs/makelinks.sh script to leave libgdp.so.0.7 around. I know this isn't the Apple-blessed way o...
08/23/2016
- 03:29 PM Bug #36 (Resolved): Error 409
- 03:27 PM Feature #42 (Closed): Need way to "clone" a log instance
- 03:26 PM Bug #29 (Resolved): Internal Server Error
- 03:11 PM Bug #39 (Resolved): Repeated "Forwarding NEW SECONDARY to node" messages, leading to "ERROR: Conn...
- 03:10 PM Bug #9 (Resolved): Router terminating connections abruptly
- 03:09 PM Bug #12 (Resolved): Connection reset for clients
- 03:09 PM Bug #32 (Resolved): Router not allowing gdplogd to connect
- 02:45 PM Bug #49 (Closed): Darwin: shared libraries have the wrong internal name
- I'm tracking down a problem where in GDP/JavaScript interface where debug flags are not preserved and ran across a pr...
- 02:12 PM Feature #48 (Closed): Scanning for corrupt indices at gdplogd startup
- Even with appropriate flushing of appends to data file and to index files (`recno=>offset` and `timstamp=>recno`), th...
- 10:51 AM Bug #47 (Closed): No sync to disk after timestamp index update
- In `gdplogd` timestamp index maintenance, there ought to be a `DB->sync()` after any update (See source:gdplogd/logd_...
08/22/2016
- 12:32 PM Bug #15 (Closed): Assertion error in Python API
- 12:32 PM Feature #31 (Resolved): Systemd debian packaging scripts for log-server
- Need some cleaning up of the adm/* scripts to make sure:
* both systemd and upstart are supported.
* dependency o... - 09:56 AM Bug #45: Reading errors (timeout, assertion failure, end of file)
- Yes, I suspect that's exactly the cause. People seem to like the idea of log auto-creation, but it has implications....
- 09:39 AM Bug #45: Reading errors (timeout, assertion failure, end of file)
- Eric Allman wrote:
> The first observation is that that log exists on both gdp-01 and gdp-04.
Could it be becaus... - 08:55 AM Bug #45: Reading errors (timeout, assertion failure, end of file)
- The first observation is that that log exists on both gdp-01 and gdp-04. It isn't clear what the semantics of that w...
- 10:48 PM Bug #45 (Closed): Reading errors (timeout, assertion failure, end of file)
- Reading on certain logs/record combinations fails with one out of three results. (on commit:5da0351b).
What causes... - 09:25 AM Bug #40 (In Progress): gcl-create calls abort: there should be a version of _gdp_gcl_decref that ...
- 09:17 AM Bug #40: gcl-create calls abort: there should be a version of _gdp_gcl_decref that does not invok...
- I can't reopen this, but it is happening again:
~~~
bash-3.2$ $PTII/vendors/gdp/gdp/apps/gcl-create -D*=10 -k non... - 09:06 AM Feature #46 (Closed): Need way of rebuilding log indices ("fsck for logs")
- Since the tidx (timestamp index) is relatively new, many logs don't have the index necessary to make "query by timest...
08/21/2016
- 01:56 PM Bug #44 (New): Router connections should check for conflicting logs
- When a router accepts a connection from another router that it has not previously seen, it should check for name clas...
- 01:53 PM Feature #42 (Resolved): Need way to "clone" a log instance
- `gcl-clone` is now in the apps directory.
08/19/2016
- 03:47 PM Feature #43 (Rejected): A default gdp router should be contacted if a local gdp router cannot be ...
- Unfortunately, building the zero-conf libavahi under Mac OS X is difficult under brew.
As a result, I added "make ... - 12:37 PM Feature #42 (Closed): Need way to "clone" a log instance
- Kaz needs a way to "clone" an existing log (including metadata) to one or more servers. Currently `gcl_create` inclu...
- 12:35 PM Bug #40 (Closed): gcl-create calls abort: there should be a version of _gdp_gcl_decref that does ...
- I'm going to close this because I can no longer reproduce it. The bug probably still persists though.
- 09:35 AM Bug #40: gcl-create calls abort: there should be a version of _gdp_gcl_decref that does not invok...
- Assigning this to myself to see if I can get a close button
- 08:54 AM Bug #41: gdp-writer segfaults under Mac OS X
- Thanks! I can no longer reproduce the bug.
- 08:31 AM Bug #41 (Resolved): gdp-writer segfaults under Mac OS X
- Resulted from failure to propagate an error if the DNS lookup failed when opening a GDP channel. Fixed in commit:f05...
08/17/2016
- 12:22 PM Bug #40: gcl-create calls abort: there should be a version of _gdp_gcl_decref that does not invok...
- I can't reproduce this, it was probably because some of the daemons were not working, or else there was a fix. I'm m...
08/15/2016
- 03:39 PM Bug #40: gcl-create calls abort: there should be a version of _gdp_gcl_decref that does not invok...
- Is gdp-03 up? I'm getting failures from both Mac OS and RHEL:
~~~
bash-3.2$ ./gcl-create -D*=10 -k none -G gdp-03...
08/14/2016
- 09:43 AM Bug #40: gcl-create calls abort: there should be a version of _gdp_gcl_decref that does not invok...
- Christopher Brooks wrote:
> I suspect that gdp-03.eecs.berkeley.edu might be down.
>
gdp-03 is working fine (ju... - 09:16 AM Bug #40 (Closed): gcl-create calls abort: there should be a version of _gdp_gcl_decref that does ...
- I suspect that gdp-03.eecs.berkeley.edu might be down.
The more immediate symptom is that gcl-create Is failing wi... - 09:34 AM Bug #41 (Closed): gdp-writer segfaults under Mac OS X
- Under Mac OS X, gdp-writer segfaults
~~~
bash-3.2$ echo "foo" | /Users/cxh/ptII/vendors/gdp/gdp/apps/gdp-writer -...
08/11/2016
- 12:43 PM Feature #37 (Closed): Add read by timestamp feature in Python API
- Commit commit:fae38f90 implements this.
08/10/2016
- 10:16 PM Bug #39 (Closed): Repeated "Forwarding NEW SECONDARY to node" messages, leading to "ERROR: Connec...
- Router goes in a pretty useless state, not allowing any clients to do useful operations. Here are the symptoms:
- ...
08/09/2016
- 03:55 PM Bug #15 (Feedback): Assertion error in Python API
- 03:53 PM Feature #13 (Closed): Flushing administrative output
- 03:52 PM Bug #14 (Closed): GDP logging
- 03:49 PM Bug #21 (Resolved): "500: internal server error" in the visualization tool.
- 03:43 PM Bug #28 (In Progress): gdplogd losing connection to the router. signature write failure: Success
- 03:40 PM Bug #30 (Closed): Zeroconf lookup hangs on reconnect attempt
- 03:37 PM Bug #35 (In Progress): EP Crypto Error
- Check if API is being properly applied.
- 03:36 PM Bug #36 (In Progress): Error 409
08/08/2016
- 08:24 AM Bug #17 (Closed): Issue with gcl-create and log subscription
- Closed: no feedback in over two weeks, presumed fixed.
- 08:22 AM Bug #38 (Resolved): Resource exhaustion after repeated router restarts
- Fixed in commit:ed4ec0a4.
- 08:22 AM Bug #38 (Closed): Resource exhaustion after repeated router restarts
- GDP clients (including gdplogd) attempt to create a new router connection if the old one goes away. The GDP library ...
08/06/2016
08/05/2016
- 04:24 PM Bug #36 (Closed): Error 409
- 2016-08-05 18:22:11.731784 -0500 mqtt-gdp-gateway: cannot log MQTT message for device/PowerBlade/c098e57000a8 (nrecs ...
- 09:17 AM Bug #35 (Closed): EP Crypto Error
- Saw this error today. Had to restart the program. Reporting for your information
EP Crypto Error: cannot clone mes...
08/04/2016
- 02:03 PM Feature #33: Mailing list for GDP Users
- Setting an email gateway is a separate issue than a mailing list. To keep track of things, I created it as a separate...
- 01:55 PM Feature #33: Mailing list for GDP Users
- It would be a pity if a email/bughandler gateway did not work.
I had bugzilla set up to handle email in 2004 on th... - 01:43 PM Feature #33: Mailing list for GDP Users
>
> BTW - I tried replying to the email that was sent from Eric's update of this issue and it looks like the issu...- 01:36 PM Feature #33: Mailing list for GDP Users
- Ok, I'm watching for all events, and the two GDP forums. If someone adds a forum, hopefully watching all events will...
- 01:21 PM Feature #33: Mailing list for GDP Users
- For Forums, select a Forum of interest and then click "Watch" in the upper right-hand corner.
For Issues, you can ... - 11:53 AM Feature #33: Mailing list for GDP Users
- This would primarily be users posting queries that might end up as being tickets. There might also be some develope...
- 11:21 AM Feature #33: Mailing list for GDP Users
- Could you clarify the nature of this mailing list? I see a few different categories:
1. Developer to developer com... - 09:49 AM Feature #33 (New): Mailing list for GDP Users
- We need a mailing list for GDP users and https://gdp.cs.berkeley.edu/redmine/projects/gdp and https://gdp.cs.berkeley...
08/03/2016
- 08:59 PM Bug #32: Router not allowing gdplogd to connect
- Eric Allman wrote:
> This looks like the Zeroconf problem I fixed this morning (nothing to do with the router). Unl... - 07:14 PM Bug #32: Router not allowing gdplogd to connect
- This looks like the Zeroconf problem I fixed this morning (nothing to do with the router). Unless you are talking ab...
08/02/2016
- 03:47 PM Bug #28: gdplogd losing connection to the router. signature write failure: Success
- Here's how to reproduce this:
1. Start `gdplogd`.
2. Use the visualization tool at http://gdp.cs.berkeley.edu:888... - 11:51 AM Bug #28: gdplogd losing connection to the router. signature write failure: Success
- The Zeroconf part of this is resolved (see issue #30). The signature write failure part is still open.
- 03:29 PM Bug #32 (Closed): Router not allowing gdplogd to connect
- Here's a full run of the gdplogd
```
[Tue Aug 2 15:23:08 PDT 2016] starting
_gdp_lib_init(NULL)
@(#)G... - 01:44 PM Feature #31 (Closed): Systemd debian packaging scripts for log-server
- 01:43 PM Bug #21: "500: internal server error" in the visualization tool.
- **Workaround:** if using with `systemd` with something like this:
```
[Service]
...
ExecStart=/usr/bin/python ... - 11:50 AM Bug #30 (Resolved): Zeroconf lookup hangs on reconnect attempt
- Fixed. The problem was that the Zeroconf code was leaving global state around that confused later calls.
- 11:48 AM Bug #30 (Closed): Zeroconf lookup hangs on reconnect attempt
- If a router goes down, the client (in this case including gdplogd) will attempt to re-connect. If it tries Zeroconf,...
Also available in: Atom