Johannes Weberhofer wrote:
I've started packaging for openSUSE, but after updating my test-system, I have recognized, that everything is very slow now. Where it took MW1.19 ~0.3 secs to render a page, I have now to wait for 12 seconds.
Is anyone else seeing this? I haven't seen any responses yet.
Is there any possibility to check in which section of the code the time is lost? If no other solution, I will next week try to see if a "standard" mediawiki from an unpacked tar is having the same issues.
Johannes
Am 25.10.2012 15:38, schrieb Mark A. Hershberger:
Johannes Weberhofer wrote:
I've started packaging for openSUSE, but after updating my test-system, I have recognized, that everything is very slow now. Where it took MW1.19 ~0.3 secs to render a page, I have now to wait for 12 seconds.
Is anyone else seeing this? I haven't seen any responses yet.
On Thu, Oct 25, 2012 at 10:00 AM, Johannes Weberhofer jweberhofer@weberhofer.at wrote:
Is there any possibility to check in which section of the code the time is lost? If no other solution, I will next week try to see if a "standard" mediawiki from an unpacked tar is having the same issues.
Johannes
http://www.mediawiki.org/wiki/Manual:How_to_debug#Profiling
Enable profiling and it will break down timings on what is taking how long.
Am 25.10.2012 17:03, schrieb OQ:
On Thu, Oct 25, 2012 at 10:00 AM, Johannes Weberhofer jweberhofer@weberhofer.at wrote:
Is there any possibility to check in which section of the code the time is lost? If no other solution, I will next week try to see if a "standard" mediawiki from an unpacked tar is having the same issues.
Johannes
http://www.mediawiki.org/wiki/Manual:How_to_debug#Profiling
Enable profiling and it will break down timings on what is taking how long.
Thank you. I've attached a logfile. Maybe someone could give me hint, what's going wrong.
Johannes
On Thu, Oct 25, 2012 at 10:24 AM, Johannes Weberhofer jweberhofer@weberhofer.at wrote:
Am 25.10.2012 17:03, schrieb OQ: Thank you. I've attached a logfile. Maybe someone could give me hint, what's going wrong.
Johannes
Forget to attach it or did mailman strip it off?
Should be able to put it on a pastebin somewhere.
Am 25.10.2012 17:34, schrieb OQ:
On Thu, Oct 25, 2012 at 10:24 AM, Johannes Weberhofer jweberhofer@weberhofer.at wrote:
Am 25.10.2012 17:03, schrieb OQ: Thank you. I've attached a logfile. Maybe someone could give me hint, what's going wrong.
Johannes
Forget to attach it or did mailman strip it off?
Should be able to put it on a pastebin somewhere.
Stripped of. You can find it here: http://www.weberhofer.at/mediawiki.log
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
On Thu, Oct 25, 2012 at 10:52 AM, Johannes Weberhofer jweberhofer@weberhofer.at wrote:
Stripped of. You can find it here: http://www.weberhofer.at/mediawiki.log
Do you have a memcached server defined that isn't actually there or is pointing to the wrong place? That's about the only thing I can think of given all the time is spent in MessageCache::load
Am 25.10.2012 18:14, schrieb OQ:
On Thu, Oct 25, 2012 at 10:52 AM, Johannes Weberhofer jweberhofer@weberhofer.at wrote:
Stripped of. You can find it here: http://www.weberhofer.at/mediawiki.log
Do you have a memcached server defined that isn't actually there or is pointing to the wrong place? That's about the only thing I can think of given all the time is spent in MessageCache::load
Thank you, for your quick response; there have been the following items in LocalSettings.php:
$wgMainCacheType = CACHE_ACCEL; $wgMemCachedServers = array();
removing those lines resolved this issue immediately. These two lines didn't have any influence in older versions of mediawiki.
Best regards, Johannes
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
On 25/10/12 18:18, Johannes Weberhofer wrote:
Thank you, for your quick response; there have been the following items in LocalSettings.php:
$wgMainCacheType = CACHE_ACCEL; $wgMemCachedServers = array();
removing those lines resolved this issue immediately. These two lines didn't have any influence in older versions of mediawiki.
That's odd. If you provided CACHE_ACCEL and you don't really have a cache accelerator which can be used, you should be getting an exception. So not something that "just slows things" (in old versions I think it was indeed silently skipped). Do you have any accelerator installed? Maybe you have one with very little space available and it was spending the time with many misses?
Am 25.10.12 19:01, schrieb Platonides:
On 25/10/12 18:18, Johannes Weberhofer wrote:
Thank you, for your quick response; there have been the following items in LocalSettings.php:
$wgMainCacheType = CACHE_ACCEL; $wgMemCachedServers = array();
removing those lines resolved this issue immediately. These two lines didn't have any influence in older versions of mediawiki.
That's odd. If you provided CACHE_ACCEL and you don't really have a cache accelerator which can be used, you should be getting an exception. So not something that "just slows things" (in old versions I think it was indeed silently skipped). Do you have any accelerator installed? Maybe you have one with very little space available and it was spending the time with many misses?
There is the php5-xcache rpm installed. Nothing else. There was no change in the system other then the version upgrade
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
On Thu, Oct 25, 2012 at 6:18 PM, Johannes Weberhofer < jweberhofer@weberhofer.at> wrote:
Am 25.10.2012 18:14, schrieb OQ:
On Thu, Oct 25, 2012 at 10:52 AM, Johannes Weberhofer
jweberhofer@weberhofer.at wrote:
Stripped of. You can find it here: http://www.weberhofer.at/** mediawiki.log http://www.weberhofer.at/mediawiki.log
Do you have a memcached server defined that isn't actually there or is pointing to the wrong place? That's about the only thing I can think of given all the time is spent in MessageCache::load
Thank you, for your quick response; there have been the following items in LocalSettings.php:
$wgMainCacheType = CACHE_ACCEL; $wgMemCachedServers = array();
removing those lines resolved this issue immediately. These two lines didn't have any influence in older versions of mediawiki.
Your log file:
CACHES: XCacheBagOStuff[main] XCacheBagOStuff[message] XCacheBagOStuff[parser] [...] MessageCache::load: Loading en... cache is empty, loading from database, loading FAILED - cache is disabled [...] MessageCache::load 1 10002.965 10002.965 98.312% 1670 ( 10002.965 - 10002.965) [1]
Shouldn't it fallback to the DB if XCache is not found? Were there changes with respect to this in 1.20?
Bryan
On 25/10/12 19:03, Bryan Tong Minh wrote:
Your log file:
CACHES: XCacheBagOStuff[main] XCacheBagOStuff[message] XCacheBagOStuff[parser] [...] MessageCache::load: Loading en... cache is empty, loading from database, loading FAILED - cache is disabled [...] MessageCache::load 1 10002.965 10002.965 98.312% 1670 ( 10002.965 - 10002.965) [1]
Shouldn't it fallback to the DB if XCache is not found? Were there changes with respect to this in 1.20?
Bryan
That's not necessarily better. The current approach of disabling the MessageCache seems good.
The strange thing is that it will note the cache error before loading the messages (and not use them). And the cache is then put into disabled mode. So I don't know why it would take so much time in MessageCache::load()
On 26 October 2012 01:07, Platonides Platonides@gmail.com wrote:
On 25/10/12 19:03, Bryan Tong Minh wrote:
Your log file:
CACHES: XCacheBagOStuff[main] XCacheBagOStuff[message] XCacheBagOStuff[parser] [...] MessageCache::load: Loading en... cache is empty, loading from database, loading FAILED - cache is disabled [...] MessageCache::load 1 10002.965 10002.965 98.312% 1670 ( 10002.965 - 10002.965) [1]
Shouldn't it fallback to the DB if XCache is not found? Were there changes with respect to this in 1.20?
Bryan
That's not necessarily better. The current approach of disabling the MessageCache seems good.
The strange thing is that it will note the cache error before loading the messages (and not use them). And the cache is then put into disabled mode. So I don't know why it would take so much time in MessageCache::load()
It tries to acquire the lock until MGS_WAIT_TIMEOUT has passed, which is 10 seconds. The lock is there to prevent multiple threads building the cache concurrently. For reasons unknown to me, there is another lock inside ->lock() ->unlock() which again fails and disables message cache.
-Niklas
I think that this issue is no show-stopper for releasing MW 1.20, as there is a clear solution for the problem.
Johannes
Am 26.10.2012 07:48, schrieb Niklas Laxström:
On 26 October 2012 01:07, Platonides Platonides@gmail.com wrote:
On 25/10/12 19:03, Bryan Tong Minh wrote:
Your log file:
CACHES: XCacheBagOStuff[main] XCacheBagOStuff[message] XCacheBagOStuff[parser] [...] MessageCache::load: Loading en... cache is empty, loading from database, loading FAILED - cache is disabled [...] MessageCache::load 1 10002.965 10002.965 98.312% 1670 ( 10002.965 - 10002.965) [1]
Shouldn't it fallback to the DB if XCache is not found? Were there changes with respect to this in 1.20?
Bryan
That's not necessarily better. The current approach of disabling the MessageCache seems good.
The strange thing is that it will note the cache error before loading the messages (and not use them). And the cache is then put into disabled mode. So I don't know why it would take so much time in MessageCache::load()
It tries to acquire the lock until MGS_WAIT_TIMEOUT has passed, which is 10 seconds. The lock is there to prevent multiple threads building the cache concurrently. For reasons unknown to me, there is another lock inside ->lock() ->unlock() which again fails and disables message cache.
-Niklas
On 26/10/12 07:48, Niklas Laxström wrote:
It tries to acquire the lock until MGS_WAIT_TIMEOUT has passed, which is 10 seconds. The lock is there to prevent multiple threads building the cache concurrently. For reasons unknown to me, there is another lock inside ->lock() ->unlock() which again fails and disables message cache.
-Niklas
However, XCache must be working for it to wait the WAIT_TIMEOUT. If there wasn't a key, the mMemc->add would have been successful (XCacheBagOStuff::set always returns true).
Maybe Johannes has XCache installed and working, but it doesn't have enough space to store the message cache, so it is always a miss, and his threads are constantly racing to rebuild it.
On Tue, Oct 30, 2012 at 10:25 PM, Platonides Platonides@gmail.com wrote:
Maybe Johannes has XCache installed and working, but it doesn't have enough space to store the message cache, so it is always a miss, and his threads are constantly racing to rebuild it.
Do we have a known minimum cache size? If so, we should check for it during installation.
Bryan
I have an install of 1.20 on openSUSE 12.1, and it's working fine.
On Thu, Oct 25, 2012 at 6:38 AM, Mark A. Hershberger mah@everybody.org wrote:
Johannes Weberhofer wrote:
I've started packaging for openSUSE, but after updating my test-system, I have recognized, that everything is very slow now. Where it took MW1.19 ~0.3 secs to render a page, I have now to wait for 12 seconds.
Is anyone else seeing this? I haven't seen any responses yet.
Any time you have "one overriding idea", and push your idea as a superior ideology, you're going to be wrong. ... The fact is, reality is complicated -- Linus Torvalds http://hexm.de/mc
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
wikitech-l@lists.wikimedia.org